🕷️ Crawler Inspector

URL Lookup

Direct Parameter Lookup

Raw Queries and Responses

1. Shard Calculation

Query:

Response:

Calculated Shard: 169 (from laksa011)

2. Crawled Status Check

Query:

curl -X POST \
  'http://laksa169.int.ahrefs:8124/' \
  -H 'Content-Type: text/plain' \
  -H 'X-ClickHouse-Database: crawler3' \
  -H 'Authorization: Basic YXBpOg==' \
  -d 'SELECT getAhrefsURLFromUnparsed(src_unparsed) AS found_url, ifNull(toUnixTimestamp(download_stamp), 0) AS crawl_time, ifNull(toUnixTimestamp(props_url_first_seen), 0) AS first_indexed_time, download_http_code AS http_code, src_unparsed AS src_unparsed, src_root_hash AS src_root_hash, history_drop_reason AS history_drop_reason, meta_title AS meta_title, meta_descriptions AS meta_descriptions, attrs_boilerpipe_text AS attrs_boilerpipe_text, attrs_markdown AS attrs_markdown, attrs_readable_markdown AS attrs_readable_markdown, meta_canonical AS meta_canonical FROM crawler3.page_info_local FINAL PREWHERE (src_root_hash, src_unparsed) IN ((getAhrefsRootHashFromUnparsed(getAhrefsUnparsedNoserviceFromURL(\'https://catboost.ai/docs/en/references/training-parameters/\')), getAhrefsUnparsedNoserviceFromURL(\'https://catboost.ai/docs/en/references/training-parameters/\'))) FORMAT JSONEachRow'

Response:

{"found_url":"https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/","crawl_time":1775997861,"first_indexed_time":1731684979,"http_code":200,"src_unparsed":"ai,catboost!\/docs\/en\/references\/training-parameters\/ s443","src_root_hash":"17435841955170310369","history_drop_reason":null,"meta_title":"Overview | CatBoost","meta_descriptions":["These parameters are for the Python package, R package and Command-line version."],"attrs_boilerpipe_text":"These parameters are for the Python package, R package and Command-line version.\nFor the Python package several parameters have aliases. For example, the\n--iterations\nparameter has the following synonyms:\nnum_boost_round\n,\nn_estimators\n,\nnum_trees\n. Simultaneous usage of different names of one parameter raises an error.\nTraining or inference on CUDA-enabled GPUs requires NVIDIA Driver of version 450.80.02 or higher.\nCommon parameters\nloss_function\nCommand-line:\n--loss-function\nAlias:\nobjective\nThe\nmetric\nto use in training. The specified value also determines the machine learning problem to solve. Some metrics support optional parameters (see the \nObjectives and metrics\nsection for details on each metric).\ncustom_metric\nCommand-line:\n--custom-metric\nMetric\nvalues to output during training. These functions are not optimized and are displayed for informational purposes only. Some metrics support optional parameters (see the \nObjectives and metrics\nsection for details on each metric).\neval_metric\nCommand-line:\n--eval-metric\nThe metric used for overfitting detection (if enabled) and best model selection (if enabled). Some metrics support optional parameters (see the \nObjectives and metrics\nsection for details on each metric).\niterations\nCommand-line:\n-i\n,\n--iterations\nAliases:\nnum_boost_round\n,\nn_estimators\n,\nnum_trees\nThe maximum number of trees that can be built when solving machine learning problems.\nlearning_rate\nCommand-line:\n-w\n,\n--learning-rate\nAlias:\neta\nThe learning rate.\nUsed for reducing the gradient step.\nrandom_seed\nCommand-line:\n-r\n,\n--random-seed\nAlias:\nrandom_state\nThe random seed used for training.\nl2_leaf_reg\nCommand-line:\n--l2-leaf-reg\n,\nl2-leaf-regularizer\nAlias:\nreg_lambda\nCoefficient at the L2 regularization term of the cost function.\nbootstrap_type\nCommand-line:\n--bootstrap-type\nBootstrap type\n. Defines the method for sampling the weights of objects.\nbagging_temperature\nCommand-line:\n--bagging-temperature\nDefines the settings of the Bayesian bootstrap. It is used by default in classification and regression modes.\nsubsample\nCommand-line:\n--subsample\nSample rate for bagging.\nsampling_frequency\nCommand-line:\n--sampling-frequency\nFrequency to sample weights and objects when building trees.\nsampling_unit\nCommand-line:\n--sampling-unit\nThe sampling scheme.\nmvs_reg\nCommand-line:\n--mvs-reg\nAffects the weight of the denominator and can be used for balancing between the importance and Bernoulli sampling (setting it to 0 implies importance sampling and to\n∞\n\\infty\n- Bernoulli).\nrandom_strength\nCommand-line:\n--random-strength\nThe amount of randomness to use for scoring splits when the tree structure is selected. Use this parameter to avoid overfitting the model.\nuse_best_model\nCommand-line:\n--use-best-model\nUse the validation dataset to identify the iteration with the optimal value of the evaluation metric.\nSave only trees up to this iteration (inclusive) in the resulting model.\nbest_model_min_trees\nCommand-line:\n--best-model-min-trees\nThe minimal number of trees that the best model should have. If set, the output model contains at least the given number of trees even if the optimal value of the evaluation metric on the validation dataset is achieved with smaller number of trees.\ndepth\nCommand-line:\n-n\n,\n--depth\nAlias:\nmax_depth\nDepth of the trees.\ngrow_policy\nCommand-line:\n--grow-policy\nThe tree growing policy. Defines how to perform greedy tree construction.\nmin_data_in_leaf\nCommand-line:\n--min-data-in-leaf\nAlias:\nmin_child_samples\nThe minimum number of training samples in a leaf. CatBoost does not search for new splits in leaves with samples count less than the specified value.\nmax_leaves\nCommand-line:\n--max-leaves\nAlias:\nnum_leaves\nThe maximum number of leafs in the resulting tree. Can be used only with the Lossguide growing policy.\nignored_features\nCommand-line:\n-I\n,\n--ignore-features\nFeature indices or names to exclude from the training. It is assumed that all passed values are feature names if at least one of the passed values can not be converted to a number or a range of numbers. Otherwise, it is assumed that all passed values are feature indices.\nSpecifics:\nNon-negative indices that do not match any features are successfully ignored. For example, if five features are defined for the objects in the dataset and this parameter is set to \n42\n, the corresponding non-existing feature is successfully ignored.\nThe identifier corresponds to the feature's index. Feature indices used in train and feature importance are numbered from 0 to\nfeatureCount – 1\n. If a file is used as \ninput data\nthen any non-feature column types are ignored when calculating these indices. For example, each row in the input file contains data in the following order:\ncat feature<\\t>label value<\\t>num feature\n. So for the row\nrock<\\t>0<\\t>42\n, the identifier for the\nrock\nfeature is 0, and for the\n42\nfeature it's 1.\nThe addition of a non-existing feature name raises an error.\none_hot_max_size\nCommand-line:\n--one-hot-max-size\nUse one-hot encoding for all categorical features with a number of different values less than or equal to the given parameter value. Ctrs are not calculated for such features.\nhas_time\nCommand-line:\n--has-time\nUse the order of objects in the input data (do not perform random permutations during the\nTransforming categorical features to numerical features\nand\nChoosing the tree structure\nstages).\nrsm\nCommand-line:\n--rsm\nAlias:\ncolsample_bylevel\nRandom subspace method. The percentage of features to use at each split selection, when features are selected over again at random.\nnan_mode\nCommand-line:\n--nan-mode\nThe method for \nprocessing missing values\nin the input dataset.\ninput_borders\nCommand-line:\n--input-borders-file\nLoad \nCustom quantization borders and missing value modes\nfrom a file (do not generate them).\noutput_borders\nCommand-line:\n--output-borders-file\nSave quantization borders for the current dataset to a file.\nfold_permutation_block\nCommand-line:\n--fold-permutation-block\nObjects in the dataset are grouped in blocks before the random permutations. This parameter defines the size of the blocks.\nleaf_estimation_method\nCommand-line:\n--leaf-estimation-method\nThe method used to calculate the values in leaves.\nleaf_estimation_iterations\nCommand-line:\n--leaf-estimation-iterations\nThis parameter regulates how many steps are done in every tree when calculating leaf values.\nleaf_estimation_backtracking\nCommand-line:\n--leaf-estimation-backtracking\nWhen the value of the\nleaf_estimation_iterations\nparameter is greater than 1, CatBoost makes several gradient or newton steps when calculating the resulting leaf values of a tree.\nfold_len_multiplier\nCommand-line:\n--fold-len-multiplier\nCoefficient for changing the length of folds.\napprox_on_full_history\nCommand-line:\n--approx-on-full-history\nThe principles for calculating the approximated values.\nclass_weights\nCommand-line:\n--class-weights\nClass weights. The values are used as multipliers for the object weights. This parameter can be used for solving binary classification and multiclassification problems.\nclass_names\nClasses names. Allows to redefine the default values when using the MultiClass and Logloss metrics.\nauto_class_weights\nCommand-line:\n--auto-class-weights\nAutomatically calculate class weights based either on the total weight or the total number of objects in each class. The values are used as multipliers for the object weights.\nSupported values:\nNone — All class weights are set to 1\nBalanced:\nC\nW\nk\n=\nm\na\nx\nc\n=\n1\nK\n(\n∑\nt\ni\n=\nc\nw\ni\n)\n∑\nt\ni\n=\nk\nw\ni\nCW_k=\\displaystyle\\frac{max_{c=1}^K(\\sum_{t_{i}=c}{w_i})}{\\sum_{t_{i}=k}{w_{i}}}\nSqrtBalanced:\nC\nW\nk\n=\nm\na\nx\nc\n=\n1\nK\n(\n∑\nt\ni\n=\nc\nw\ni\n)\n∑\nt\ni\n=\nk\nw\ni\nCW_k=\\sqrt{\\displaystyle\\frac{max_{c=1}^K(\\sum_{t_i=c}{w_i})}{\\sum_{t_i=k}{w_i}}}\nscale_pos_weight\nThe weight for class 1 in binary classification. The value is used as a multiplier for the weights of objects from class 1.\nboosting_type\nCommand-line:\n--boosting-type\nBoosting scheme.\nboost_from_average\nCommand-line:\n--boost-from-average\nInitialize approximate values by best constant value for the specified loss function.\nlangevin\nCommand-line:\n--langevin\nEnables the Stochastic Gradient Langevin Boosting mode.\ndiffusion_temperature\nCommand-line:\n--diffusion-temperature\nThe diffusion temperature of the Stochastic Gradient Langevin Boosting mode.\nposterior_sampling\nCommand-line:\n--posterior-sampling\nIf this parameter is set several options are specified as follows and model parameters are checked to obtain uncertainty predictions with good theoretical properties.\nallow_const_label\nCommand-line:\n--allow-const-label\nUse it to train models with datasets that have equal label values for all objects.\nscore_function\nCommand-line:\n--score-function\nThe\nscore type\nused to select the next split during the tree construction.\nmonotone_constraints\nCommand-line:\n--monotone-constraints\nImpose monotonic constraints on numerical features.\nPossible values:\n1\n — Increasing constraint on the feature. The algorithm forces the model to be a non-decreasing function of this features.\n-1\n — Decreasing constraint on the feature. The algorithm forces the model to be a non-increasing function of this features.\n0\n — constraints are disabled.\nfeature_weights\nCommand-line:\n--feature-weights\nPer-feature multiplication weights used when choosing the best split. The score of each candidate is multiplied by the weights of features from the current split.\nNon-negative float values are supported for each weight.\nSupported formats for setting the value of this parameter:\nfirst_feature_use_penalties\nCommand-line:\n--first-feature-use-penalties\nPer-feature penalties for the first occurrence of the feature in the model. The given value is subtracted from the score if the current candidate is the first one to include the feature in the model.\nRefer to the\nPer-object and per-feature penalties\nsection for details on applying different score penalties.\nNon-negative float values are supported for each penalty.\nfixed_binary_splits\nCommand-line:\n--fixed-binary-splits\nA list of indices of binary features to put at the top of each tree.\npenalties_coefficient\nCommand-line:\n--penalties-coefficient\nA single-value common coefficient to multiply all penalties.\nper_object_feature_penalties\nCommand-line:\n--per-object-feature-penalties\nPer-object penalties for the first use of the feature for the object. The given value is multiplied by the number of objects that are divided by the current split and use the feature for the first time.\nRefer to the\nPer-object and per-feature penalties\nsection for details on applying different score penalties.\nNon-negative float values are supported for each penalty.\nmodel_shrink_rate\nCommand-line:\n--model-shrink-rate\nThe constant used to calculate the coefficient for multiplying the model on each iteration.\nmodel_shrink_mode\nCommand-line:\nmodel_shrink_mode\nDetermines how the actual model shrinkage coefficient is calculated at each iteration.\nCTR settings\nsimple_ctr\nQuantization settings for simple \ncategorical features\n. Use this parameter to specify the principles for defining the class of the object for regression tasks. By default, it is considered that an object belongs to the positive class if its' label value is greater than the median of all label values of the dataset.\ncombinations_ctr\nQuantization settings for combinations of \ncategorical features\n.\nper_feature_ctr\nPer-feature quantization settings for \ncategorical features\n.\nctr_target_border_count\nThe maximum number of borders to use in target quantization for categorical features that need it. Allowed values are integers from 1 to 255 inclusively.\ncounter_calc_method\nThe method for calculating the Counter CTR type.\nmax_ctr_complexity\nThe maximum number of features that can be combined.\nctr_leaf_count_limit\nThe maximum number of leaves with categorical features. If the quantity exceeds the specified value a part of leaves is discarded.\nstore_all_simple_ctr\nIgnore categorical features, which are not used in feature combinations, when choosing candidates for exclusion.\nfinal_ctr_computation_mode\nFinal CTR computation mode.\nInput file settings\n-f, --learn-set\nThe path to the input file  that contains the dataset description.\n-t, --test-set\nA comma-separated list of input files that contain the validation dataset description (the format must be the same as used in the training dataset).\n--cd, --column-description\nThe path to the input file  that contains the\ncolumns description\n.\n--learn-pairs\nThe path to the input file that contains the \npairs description\nfor the training dataset.\n--test-pairs\nThe path to the input file that contains the \npairs description\nfor the validation dataset.\n--learn-group-weights\nThe path to the input file that contains the weights of groups. Refer to the \nGroup weights\nsection for format details.\n--test-group-weights\nThe path to the input file that contains the weights of groups for the validation dataset.\n--learn-baseline\nThe path to the input file that contains baseline values for the training dataset.\n--test-baseline\nThe path to the input file that contains baseline values for the validation dataset.\n--delimiter\nThe delimiter character used to separate the data in the dataset description input file.\nRead the column names from the first line of the dataset description file if this parameter is set.\n--params-files\nThe path to the input JSON file that contains the training parameters, for example:\n--nan-mode\nThe method for\nprocessing missing values\nin the input dataset.\nMulticlassification settings\nclasses_count\nCommand-line:\n--classes-count\nThe upper limit for the numeric class label. Defines the number of classes for multiclassification.\n--class-names\nThis parameter is only for Command-line.\nClasses names. Allows to redefine the default values when using the MultiClass and Logloss metrics.\nOutput settings\nlogging_level\nCommand line:\n--logging-level\nThe logging level to output to stdout.\nmetric_period\nCommand line:\n--metric-period\nThe frequency of iterations to calculate the values of \nobjectives and metrics\n.\nThe usage of this parameter speeds up the training.\nverbose\nCommand line:\n--verbose\nAlias:\nverbose_eval\nThe purpose of this parameter depends on the type of the given value:\ntrain_dir\nCommand line:\n--train-dir\nThe directory for storing the files generated during training.\nmodel_size_reg\nCommand line:\n--model-size-reg\nThe model size regularization coefficient. The larger the value, the smaller the model size. Refer to the\nModel size regularization coefficient\nsection for details.\nThis regularization is needed only for models with categorical features (other models are small).\nallow_writing_files\nAllow to write analytical and snapshot files during training.\nsave_snapshot\nEnable snapshotting for \nrestoring the training progress after an interruption\n.\nsnapshot_file\nThe name of the file to save the training progress information in. This file is used for \nrecovering training after an interruption\n.\nsnapshot_interval\nThe interval between saving snapshots in seconds.\nroc_file\nThe name of the\noutput file\nto save the ROC curve points to.\nOverfitting detection settings\nearly_stopping_rounds\nSets the overfitting detector type to Iter and stops the training after the specified number of iterations since the iteration with the optimal metric value.\nod_type\nCommand-line:\n--od-type\nThe type of the overfitting detector to use.\nod_pval\nCommand-line:\n--od-pval\nThe threshold for the IncToDec\noverfitting detector\ntype.\nod_wait\nCommand-line:\n--od-wait\nThe number of iterations to continue the training after the iteration with the optimal metric value.\nPerformance settings\nthread_count\nCommand-line:\n-T\n,\n--thread-count\nThe number of threads to use during the training.\nused_ram_limit\nCommand-line:\n--used-ram-limit\nAttempt to limit the amount of used CPU RAM.\ngpu_ram_part\nCommand-line:\n--gpu-ram-part\nHow much of the GPU RAM to use for training.\npinned_memory_size\nCommand-line:\n--pinned-memory-size\nHow much pinned (page-locked) CPU RAM to use per GPU.\ngpu_cat_features_storage\nCommand-line:\n--gpu-cat-features-storage\nThe method for storing the categorical features' values.\ndata_partition\nCommand-line:\n--data-partition\nThe method for splitting the input dataset between multiple workers.\nProcessing unit settings\ntask_type\nCommand line:\n--task-type\nThe processing unit type to use for training.\ndevices\nCommand line:\n--devices\nIDs of the GPU devices to use for training (indices are zero-based).\nQuantization settings\ntarget_border\nCommand-line:\n--target-border\nIf set, defines the border for converting target values to 0 and 1.\nborder_count\nCommand-line:\n-x\n,\n--border-count\nAlias:\nmax_bin\nThe number of splits for numerical features. Allowed values are integers from 1 to 65535 inclusively.\nfeature_border_type\nCommand-line:\n--feature-border-type\nThe\nquantization mode\nfor numerical features.\nper_float_feature_quantization\nCommand-line:\n--per-float-feature-quantization\nThe quantization description for the specified feature or list of features.\nText processing parameters\nThese parameters are only for the Python package and Command-line version.\ntokenizers\nCommand-line:\n--tokenizers\nTokenizers used to preprocess Text type feature columns before creating the dictionary.\ndictionaries\nCommand-line:\n--dictionaries\nDictionaries used to preprocess Text type feature columns.\nFormat:\nfeature_calcers\nCommand-line:\n--feature-calcers\nFeature calcers used to calculate new features based on preprocessed Text type feature columns.\nFormat:\ntext_processing\nCommand-line:\n--text-processing\nA JSON specification of tokenizers, dictionaries and feature calcers, which determine how text features are converted into a list of float features.\nExample\nRefer to the description of the following parameters for details on supported values:\nVisualization settings\nThese parameters are only for the Python package.\nname\nThe experiment name to display in\nvisualization tools\n.","attrs_markdown":"[![Logo icon](https:\/\/yastatic.net\/s3\/locdoc\/daas-static\/catboost\/71b237a322eec6f2889af0dae2a9c549.svg)](https:\/\/catboost.ai\/ \"CatBoost\")\n\n- Installation\n  - [Overview](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/installation)\n  - Python package installation\n  - CatBoost for Apache Spark installation\n  - R package installation\n  - Command-line version binary\n  - Build from source\n- Key Features\n- Training parameters\n  - [Overview](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/)\n  - [Common parameters](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common)\n  - [CTR settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/ctr)\n  - [Input settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/input)\n  - [Multiclassification settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/multiclassification)\n  - [Output settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/output)\n  - [Overfitting detection settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/overfitting-detection)\n  - [Performance settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/performance)\n  - [Processing unit settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/processing-unit)\n  - [Quantization settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/quantization)\n  - [Text processing parameters](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/text-processing)\n  - [Visualization settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/visualization)\n- Python package\n- CatBoost for Apache Spark\n- R package\n- Command-line version\n- Applying models\n- Objectives and metrics\n- Model analysis\n- Data format description\n- [Parameter tuning](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/parameter-tuning)\n- [Speeding up the training](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/speed-up-training)\n- Data visualization\n- Algorithm details\n- [FAQ](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/faq)\n- Educational materials\n- [Development and contributions](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/development-and-contributions)\n- [Contacts](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/contacts)\n\nOverview\n\n## In this article:\n- [Common parameters](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#common-parameters)\n- [loss\\_function](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#loss_function)\n- [custom\\_metric](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#custom_metric)\n- [eval\\_metric](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#eval_metric)\n- [iterations](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#iterations)\n- [learning\\_rate](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#learning_rate)\n- [random\\_seed](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#random_seed)\n- [l2\\_leaf\\_reg](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#l2_leaf_reg)\n- [bootstrap\\_type](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#bootstrap_type)\n- [bagging\\_temperature](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#bagging_temperature)\n- [subsample](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#subsample)\n- [sampling\\_frequency](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#sampling_frequency)\n- [sampling\\_unit](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#sampling_unit)\n- [mvs\\_reg](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#mvs_reg)\n- [random\\_strength](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#random_strength)\n- [use\\_best\\_model](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#use_best_model)\n- [best\\_model\\_min\\_trees](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#best_model_min_trees)\n- [depth](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#depth)\n- [grow\\_policy](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#grow_policy)\n- [min\\_data\\_in\\_leaf](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#min_data_in_leaf)\n- [max\\_leaves](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#max_leaves)\n- [ignored\\_features](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#ignored_features)\n- [one\\_hot\\_max\\_size](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#one_hot_max_size)\n- [has\\_time](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#has_time)\n- [rsm](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#rsm)\n- [nan\\_mode](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#nan_mode)\n- [input\\_borders](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#input_borders)\n- [output\\_borders](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#output_borders)\n- [fold\\_permutation\\_block](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#fold_permutation_block)\n- [leaf\\_estimation\\_method](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#leaf_estimation_method)\n- [leaf\\_estimation\\_iterations](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#leaf_estimation_iterations)\n- [leaf\\_estimation\\_backtracking](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#leaf_estimation_backtracking)\n- [fold\\_len\\_multiplier](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#fold_len_multiplier)\n- [approx\\_on\\_full\\_history](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#approx_on_full_history)\n- [class\\_weights](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#class_weights)\n- [class\\_names](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#class_names)\n- [auto\\_class\\_weights](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#auto_class_weights)\n- [scale\\_pos\\_weight](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#scale_pos_weight)\n- [boosting\\_type](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#boosting_type)\n- [boost\\_from\\_average](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#boost_from_average)\n- [langevin](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#langevin)\n- [diffusion\\_temperature](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#diffusion_temperature)\n- [posterior\\_sampling](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#posterior_sampling)\n- [allow\\_const\\_label](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#allow_const_label)\n- [score\\_function](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#score_function)\n- [monotone\\_constraints](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#monotone_constraints)\n- [feature\\_weights](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#feature_weights)\n- [first\\_feature\\_use\\_penalties](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#first_feature_use_penalties)\n- [fixed\\_binary\\_splits](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#fixed_binary_splits)\n- [penalties\\_coefficient](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#penalties_coefficient)\n- [per\\_object\\_feature\\_penalties](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#per_object_feature_penalties)\n- [model\\_shrink\\_rate](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#model_shrink_rate)\n- [model\\_shrink\\_mode](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#model_shrink_mode)\n- [CTR settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#ctr-settings)\n- [simple\\_ctr](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#simple_ctr)\n- [combinations\\_ctr](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#combinations_ctr)\n- [per\\_feature\\_ctr](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#per_feature_ctr)\n- [ctr\\_target\\_border\\_count](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#ctr_target_border_count)\n- [counter\\_calc\\_method](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#counter_calc_method)\n- [max\\_ctr\\_complexity](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#max_ctr_complexity)\n- [ctr\\_leaf\\_count\\_limit](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#ctr_leaf_count_limit)\n- [store\\_all\\_simple\\_ctr](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#store_all_simple_ctr)\n- [final\\_ctr\\_computation\\_mode](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#final_ctr_computation_mode)\n- [Input file settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#input-file-settings)\n- [\\-f, --learn-set](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#f,-learn-set)\n- [\\-t, --test-set](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#t,-test-set)\n- [\\--cd, --column-description](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#cd,-column-description)\n- [\\--learn-pairs](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#learn-pairs)\n- [\\--test-pairs](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#test-pairs)\n- [\\--learn-group-weights](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#learn-group-weights)\n- [\\--test-group-weights](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#test-group-weights)\n- [\\--learn-baseline](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#learn-baseline)\n- [\\--test-baseline](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#test-baseline)\n- [\\--delimiter](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#delimiter)\n- [\\--has-header](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#has-header)\n- [\\--params-files](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#params-files)\n- [\\--nan-mode](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#nan-mode)\n- [Multiclassification settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#multiclassification-settings)\n- [classes\\_count](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#classes_count)\n- [\\--class-names](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#class-names)\n- [Output settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#output-settings)\n- [logging\\_level](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#logging_level)\n- [metric\\_period](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#metric_period)\n- [verbose](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#verbose)\n- [train\\_dir](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#train_dir)\n- [model\\_size\\_reg](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#model_size_reg)\n- [allow\\_writing\\_files](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#allow_writing_files)\n- [save\\_snapshot](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#save_snapshot)\n- [snapshot\\_file](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#snapshot_file)\n- [snapshot\\_interval](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#snapshot_interval)\n- [roc\\_file](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#roc_file)\n- [Overfitting detection settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#overfitting-detection-settings)\n- [early\\_stopping\\_rounds](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#early_stopping_rounds)\n- [od\\_type](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#od_type)\n- [od\\_pval](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#od_pval)\n- [od\\_wait](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#od_wait)\n- [Performance settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#performance-settings)\n- [thread\\_count](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#thread_count)\n- [used\\_ram\\_limit](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#used_ram_limit)\n- [gpu\\_ram\\_part](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#gpu_ram_part)\n- [pinned\\_memory\\_size](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#pinned_memory_size)\n- [gpu\\_cat\\_features\\_storage](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#gpu_cat_features_storage)\n- [data\\_partition](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#data_partition)\n- [Processing unit settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#processing-unit-settings)\n- [task\\_type](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#task_type)\n- [devices](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#devices)\n- [Quantization settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#quantization-settings)\n- [target\\_border](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#target_border)\n- [border\\_count](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#border_count)\n- [feature\\_border\\_type](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#feature_border_type)\n- [per\\_float\\_feature\\_quantization](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#per_float_feature_quantization)\n- [Text processing parameters](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#text-processing-parameters)\n- [tokenizers](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#tokenizers)\n- [dictionaries](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#dictionaries)\n- [feature\\_calcers](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#feature_calcers)\n- [text\\_processing](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#text_processing)\n- [Visualization settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#visualization-settings)\n- [name](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#name)\n\n1. Training parameters\n2. Overview\n\n# Overview\n- [Common parameters](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#common-parameters)\n  - [loss\\_function](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#loss_function)\n  - [custom\\_metric](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#custom_metric)\n  - [eval\\_metric](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#eval_metric)\n  - [iterations](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#iterations)\n  - [learning\\_rate](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#learning_rate)\n  - [random\\_seed](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#random_seed)\n  - [l2\\_leaf\\_reg](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#l2_leaf_reg)\n  - [bootstrap\\_type](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#bootstrap_type)\n  - [bagging\\_temperature](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#bagging_temperature)\n  - [subsample](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#subsample)\n  - [sampling\\_frequency](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#sampling_frequency)\n  - [sampling\\_unit](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#sampling_unit)\n  - [mvs\\_reg](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#mvs_reg)\n  - [random\\_strength](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#random_strength)\n  - [use\\_best\\_model](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#use_best_model)\n  - [best\\_model\\_min\\_trees](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#best_model_min_trees)\n  - [depth](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#depth)\n  - [grow\\_policy](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#grow_policy)\n  - [min\\_data\\_in\\_leaf](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#min_data_in_leaf)\n  - [max\\_leaves](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#max_leaves)\n  - [ignored\\_features](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#ignored_features)\n  - [one\\_hot\\_max\\_size](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#one_hot_max_size)\n  - [has\\_time](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#has_time)\n  - [rsm](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#rsm)\n  - [nan\\_mode](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#nan_mode)\n  - [input\\_borders](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#input_borders)\n  - [output\\_borders](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#output_borders)\n  - [fold\\_permutation\\_block](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#fold_permutation_block)\n  - [leaf\\_estimation\\_method](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#leaf_estimation_method)\n  - [leaf\\_estimation\\_iterations](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#leaf_estimation_iterations)\n  - [leaf\\_estimation\\_backtracking](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#leaf_estimation_backtracking)\n  - [fold\\_len\\_multiplier](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#fold_len_multiplier)\n  - [approx\\_on\\_full\\_history](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#approx_on_full_history)\n  - [class\\_weights](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#class_weights)\n  - [class\\_names](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#class_names)\n  - [auto\\_class\\_weights](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#auto_class_weights)\n  - [scale\\_pos\\_weight](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#scale_pos_weight)\n  - [boosting\\_type](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#boosting_type)\n  - [boost\\_from\\_average](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#boost_from_average)\n  - [langevin](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#langevin)\n  - [diffusion\\_temperature](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#diffusion_temperature)\n  - [posterior\\_sampling](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#posterior_sampling)\n  - [allow\\_const\\_label](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#allow_const_label)\n  - [score\\_function](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#score_function)\n  - [monotone\\_constraints](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#monotone_constraints)\n  - [feature\\_weights](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#feature_weights)\n  - [first\\_feature\\_use\\_penalties](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#first_feature_use_penalties)\n  - [fixed\\_binary\\_splits](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#fixed_binary_splits)\n  - [penalties\\_coefficient](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#penalties_coefficient)\n  - [per\\_object\\_feature\\_penalties](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#per_object_feature_penalties)\n  - [model\\_shrink\\_rate](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#model_shrink_rate)\n  - [model\\_shrink\\_mode](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#model_shrink_mode)\n- [CTR settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#ctr-settings)\n  - [simple\\_ctr](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#simple_ctr)\n  - [combinations\\_ctr](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#combinations_ctr)\n  - [per\\_feature\\_ctr](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#per_feature_ctr)\n  - [ctr\\_target\\_border\\_count](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#ctr_target_border_count)\n  - [counter\\_calc\\_method](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#counter_calc_method)\n  - [max\\_ctr\\_complexity](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#max_ctr_complexity)\n  - [ctr\\_leaf\\_count\\_limit](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#ctr_leaf_count_limit)\n  - [store\\_all\\_simple\\_ctr](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#store_all_simple_ctr)\n  - [final\\_ctr\\_computation\\_mode](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#final_ctr_computation_mode)\n- [Input file settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#input-file-settings)\n  - [\\-f, --learn-set](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#f,-learn-set)\n  - [\\-t, --test-set](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#t,-test-set)\n  - [\\--cd, --column-description](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#cd,-column-description)\n  - [\\--learn-pairs](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#learn-pairs)\n  - [\\--test-pairs](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#test-pairs)\n  - [\\--learn-group-weights](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#learn-group-weights)\n  - [\\--test-group-weights](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#test-group-weights)\n  - [\\--learn-baseline](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#learn-baseline)\n  - [\\--test-baseline](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#test-baseline)\n  - [\\--delimiter](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#delimiter)\n  - [\\--has-header](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#has-header)\n  - [\\--params-files](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#params-files)\n  - [\\--nan-mode](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#nan-mode)\n- [Multiclassification settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#multiclassification-settings)\n  - [classes\\_count](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#classes_count)\n  - [\\--class-names](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#class-names)\n- [Output settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#output-settings)\n  - [logging\\_level](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#logging_level)\n  - [metric\\_period](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#metric_period)\n  - [verbose](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#verbose)\n  - [train\\_dir](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#train_dir)\n  - [model\\_size\\_reg](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#model_size_reg)\n  - [allow\\_writing\\_files](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#allow_writing_files)\n  - [save\\_snapshot](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#save_snapshot)\n  - [snapshot\\_file](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#snapshot_file)\n  - [snapshot\\_interval](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#snapshot_interval)\n  - [roc\\_file](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#roc_file)\n- [Overfitting detection settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#overfitting-detection-settings)\n  - [early\\_stopping\\_rounds](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#early_stopping_rounds)\n  - [od\\_type](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#od_type)\n  - [od\\_pval](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#od_pval)\n  - [od\\_wait](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#od_wait)\n- [Performance settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#performance-settings)\n  - [thread\\_count](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#thread_count)\n  - [used\\_ram\\_limit](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#used_ram_limit)\n  - [gpu\\_ram\\_part](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#gpu_ram_part)\n  - [pinned\\_memory\\_size](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#pinned_memory_size)\n  - [gpu\\_cat\\_features\\_storage](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#gpu_cat_features_storage)\n  - [data\\_partition](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#data_partition)\n- [Processing unit settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#processing-unit-settings)\n  - [task\\_type](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#task_type)\n  - [devices](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#devices)\n- [Quantization settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#quantization-settings)\n  - [target\\_border](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#target_border)\n  - [border\\_count](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#border_count)\n  - [feature\\_border\\_type](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#feature_border_type)\n  - [per\\_float\\_feature\\_quantization](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#per_float_feature_quantization)\n- [Text processing parameters](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#text-processing-parameters)\n  - [tokenizers](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#tokenizers)\n  - [dictionaries](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#dictionaries)\n  - [feature\\_calcers](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#feature_calcers)\n  - [text\\_processing](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#text_processing)\n- [Visualization settings](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#visualization-settings)\n  - [name](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#name)\n\nThese parameters are for the Python package, R package and Command-line version.\n\nFor the Python package several parameters have aliases. For example, the `--iterations` parameter has the following synonyms: `num_boost_round`, `n_estimators`, `num_trees`. Simultaneous usage of different names of one parameter raises an error.\n\nTraining or inference on CUDA-enabled GPUs requires NVIDIA Driver of version 450.80.02 or higher.\n\n## Common parameters\n### [loss\\_function](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#loss_function)\nCommand-line: `--loss-function`\n\n*Alias:* `objective`\n\nThe [metric](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/loss-functions) to use in training. The specified value also determines the machine learning problem to solve. Some metrics support optional parameters (see the [Objectives and metrics](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/loss-functions) section for details on each metric).\n\n### [custom\\_metric](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#custom_metric)\nCommand-line: `--custom-metric`\n\n[Metric](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/loss-functions) values to output during training. These functions are not optimized and are displayed for informational purposes only. Some metrics support optional parameters (see the [Objectives and metrics](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/loss-functions) section for details on each metric).\n\n### [eval\\_metric](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#eval_metric)\nCommand-line: `--eval-metric`\n\nThe metric used for overfitting detection (if enabled) and best model selection (if enabled). Some metrics support optional parameters (see the [Objectives and metrics](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/loss-functions) section for details on each metric).\n\n### [iterations](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#iterations)\nCommand-line: `-i`, `--iterations`\n\n*Aliases:* `num_boost_round`, `n_estimators`, `num_trees`\n\nThe maximum number of trees that can be built when solving machine learning problems.\n\n### [learning\\_rate](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#learning_rate)\nCommand-line: `-w`, `--learning-rate`\n\n*Alias:* `eta`\n\nThe learning rate.\n\nUsed for reducing the gradient step.\n\n### [random\\_seed](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#random_seed)\nCommand-line: `-r`, `--random-seed`\n\n*Alias:*`random_state`\n\nThe random seed used for training.\n\n### [l2\\_leaf\\_reg](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#l2_leaf_reg)\nCommand-line: `--l2-leaf-reg`, `l2-leaf-regularizer`\n\n*Alias:* `reg_lambda`\n\nCoefficient at the L2 regularization term of the cost function.\n\n### [bootstrap\\_type](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#bootstrap_type)\nCommand-line: `--bootstrap-type`\n\n[Bootstrap type](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/algorithm-main-stages_bootstrap-options). Defines the method for sampling the weights of objects.\n\n### [bagging\\_temperature](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#bagging_temperature)\nCommand-line: `--bagging-temperature`\n\nDefines the settings of the Bayesian bootstrap. It is used by default in classification and regression modes.\n\n### [subsample](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#subsample)\nCommand-line: `--subsample`\n\nSample rate for bagging.\n\n### [sampling\\_frequency](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#sampling_frequency)\nCommand-line: `--sampling-frequency`\n\nFrequency to sample weights and objects when building trees.\n\n### [sampling\\_unit](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#sampling_unit)\nCommand-line: `--sampling-unit`\n\nThe sampling scheme.\n\n### [mvs\\_reg](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#mvs_reg)\nCommand-line: `--mvs-reg`\n\nAffects the weight of the denominator and can be used for balancing between the importance and Bernoulli sampling (setting it to 0 implies importance sampling and to ∞ \\\\infty ∞ - Bernoulli).\n\n### [random\\_strength](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#random_strength)\nCommand-line: `--random-strength`\n\nThe amount of randomness to use for scoring splits when the tree structure is selected. Use this parameter to avoid overfitting the model.\n\n### [use\\_best\\_model](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#use_best_model)\nCommand-line: `--use-best-model`\n\nUse the validation dataset to identify the iteration with the optimal value of the evaluation metric.  \n Save only trees up to this iteration (inclusive) in the resulting model.\n\n### [best\\_model\\_min\\_trees](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#best_model_min_trees)\nCommand-line: `--best-model-min-trees`\n\nThe minimal number of trees that the best model should have. If set, the output model contains at least the given number of trees even if the optimal value of the evaluation metric on the validation dataset is achieved with smaller number of trees.\n\n### [depth](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#depth)\nCommand-line: `-n`, `--depth`\n\n*Alias:* `max_depth`\n\nDepth of the trees.\n\n### [grow\\_policy](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#grow_policy)\nCommand-line: `--grow-policy`\n\nThe tree growing policy. Defines how to perform greedy tree construction.\n\n### [min\\_data\\_in\\_leaf](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#min_data_in_leaf)\nCommand-line: `--min-data-in-leaf`\n\n*Alias:* `min_child_samples`\n\nThe minimum number of training samples in a leaf. CatBoost does not search for new splits in leaves with samples count less than the specified value.\n\n### [max\\_leaves](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#max_leavescommon.md#)\nCommand-line: `--max-leaves`\n\n*Alias:*`num_leaves`\n\nThe maximum number of leafs in the resulting tree. Can be used only with the Lossguide growing policy.\n\n### [ignored\\_features](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#ignored_features)\nCommand-line: `-I`, `--ignore-features`\n\nFeature indices or names to exclude from the training. It is assumed that all passed values are feature names if at least one of the passed values can not be converted to a number or a range of numbers. Otherwise, it is assumed that all passed values are feature indices.\n\nSpecifics:\n\n- Non-negative indices that do not match any features are successfully ignored. For example, if five features are defined for the objects in the dataset and this parameter is set to \"42\", the corresponding non-existing feature is successfully ignored.\n- The identifier corresponds to the feature's index. Feature indices used in train and feature importance are numbered from 0 to `featureCount – 1`. If a file is used as [input data](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/input-data) then any non-feature column types are ignored when calculating these indices. For example, each row in the input file contains data in the following order: `cat feature<\\t>label value<\\t>num feature`. So for the row `rock<\\t>0<\\t>42`, the identifier for the \"rock\" feature is 0, and for the \"42\" feature it's 1.\n- The addition of a non-existing feature name raises an error.\n### [one\\_hot\\_max\\_size](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#one_hot_max_size)\nCommand-line: `--one-hot-max-size`\n\nUse one-hot encoding for all categorical features with a number of different values less than or equal to the given parameter value. Ctrs are not calculated for such features.\n\n### [has\\_time](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#has_time)\nCommand-line: `--has-time`\n\nUse the order of objects in the input data (do not perform random permutations during the [Transforming categorical features to numerical features](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/algorithm-main-stages_cat-to-numberic) and [Choosing the tree structure](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/algorithm-main-stages_choose-tree-structure) stages).\n\n### [rsm](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#rsm)\nCommand-line: `--rsm`\n\n*Alias:*`colsample_bylevel`\n\nRandom subspace method. The percentage of features to use at each split selection, when features are selected over again at random.\n\n### [nan\\_mode](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#nan_mode)\nCommand-line: `--nan-mode`\n\nThe method for [processing missing values](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/algorithm-missing-values-processing) in the input dataset.\n\n### [input\\_borders](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#input_borders)\nCommand-line: `--input-borders-file`\n\nLoad [Custom quantization borders and missing value modes](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/input-data_custom-borders) from a file (do not generate them).\n\n### [output\\_borders](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#output_borders)\nCommand-line: `--output-borders-file`\n\nSave quantization borders for the current dataset to a file.\n\n### [fold\\_permutation\\_block](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#fold_permutation_block)\nCommand-line: `--fold-permutation-block`\n\nObjects in the dataset are grouped in blocks before the random permutations. This parameter defines the size of the blocks.\n\n### [leaf\\_estimation\\_method](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#leaf_estimation_method)\nCommand-line: `--leaf-estimation-method`\n\nThe method used to calculate the values in leaves.\n\n### [leaf\\_estimation\\_iterations](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#leaf_estimation_iterations)\nCommand-line: `--leaf-estimation-iterations`\n\nThis parameter regulates how many steps are done in every tree when calculating leaf values.\n\n### [leaf\\_estimation\\_backtracking](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#leaf_estimation_backtracking)\nCommand-line: `--leaf-estimation-backtracking`\n\nWhen the value of the `leaf_estimation_iterations` parameter is greater than 1, CatBoost makes several gradient or newton steps when calculating the resulting leaf values of a tree.\n\n### [fold\\_len\\_multiplier](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#fold_len_multiplier)\nCommand-line: `--fold-len-multiplier`\n\nCoefficient for changing the length of folds.\n\n### [approx\\_on\\_full\\_history](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#approx_on_full_history)\nCommand-line:`--approx-on-full-history`\n\nThe principles for calculating the approximated values.\n\n### [class\\_weights](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#class_weights)\nCommand-line: `--class-weights`\n\nClass weights. The values are used as multipliers for the object weights. This parameter can be used for solving binary classification and multiclassification problems.\n\n### [class\\_names](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#class_names)\nClasses names. Allows to redefine the default values when using the MultiClass and Logloss metrics.\n\n### [auto\\_class\\_weights](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#auto_class_weights)\nCommand-line: `--auto-class-weights`\n\nAutomatically calculate class weights based either on the total weight or the total number of objects in each class. The values are used as multipliers for the object weights.\n\nSupported values:\n\n- None — All class weights are set to 1\n- Balanced:\n  C W k \\= m a x c \\= 1 K ( ∑ t i \\= c w i ) ∑ t i \\= k w i CW\\_k=\\\\displaystyle\\\\frac{max\\_{c=1}^K(\\\\sum\\_{t\\_{i}=c}{w\\_i})}{\\\\sum\\_{t\\_{i}=k}{w\\_{i}}} CWk\\=∑ti\\=kwimaxc\\=1K(∑ti\\=cwi)\n- SqrtBalanced:\n  C W k \\= m a x c \\= 1 K ( ∑ t i \\= c w i ) ∑ t i \\= k w i CW\\_k=\\\\sqrt{\\\\displaystyle\\\\frac{max\\_{c=1}^K(\\\\sum\\_{t\\_i=c}{w\\_i})}{\\\\sum\\_{t\\_i=k}{w\\_i}}} CWk\\= ∑ti\\=kwimaxc\\=1K(∑ti\\=cwi) \n### [scale\\_pos\\_weight](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#scale_pos_weight)\nThe weight for class 1 in binary classification. The value is used as a multiplier for the weights of objects from class 1.\n\n### [boosting\\_type](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#boosting_type)\nCommand-line: `--boosting-type`\n\nBoosting scheme.\n\n### [boost\\_from\\_average](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#boost_from_average)\nCommand-line: `--boost-from-average`\n\nInitialize approximate values by best constant value for the specified loss function.\n\n### [langevin](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#langevin)\nCommand-line: `--langevin`\n\nEnables the Stochastic Gradient Langevin Boosting mode.\n\n### [diffusion\\_temperature](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#diffusion_temperature)\nCommand-line: `--diffusion-temperature`\n\nThe diffusion temperature of the Stochastic Gradient Langevin Boosting mode.\n\n### [posterior\\_sampling](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#posterior_sampling)\nCommand-line: `--posterior-sampling`\n\nIf this parameter is set several options are specified as follows and model parameters are checked to obtain uncertainty predictions with good theoretical properties.\n\n### [allow\\_const\\_label](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#allow_const_label)\nCommand-line: `--allow-const-label`\n\nUse it to train models with datasets that have equal label values for all objects.\n\n### [score\\_function](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#score_function)\nCommand-line: `--score-function`\n\nThe [score type](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/algorithm-score-functions) used to select the next split during the tree construction.\n\n### [monotone\\_constraints](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#monotone_constraints)\nCommand-line: `--monotone-constraints`\n\nImpose monotonic constraints on numerical features.\n\nPossible values:\n\n- \"1\" — Increasing constraint on the feature. The algorithm forces the model to be a non-decreasing function of this features.\n- \"\\-1\" — Decreasing constraint on the feature. The algorithm forces the model to be a non-increasing function of this features.\n- \"0\" — constraints are disabled.\n### [feature\\_weights](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#feature_weights)\nCommand-line: `--feature-weights`\n\nPer-feature multiplication weights used when choosing the best split. The score of each candidate is multiplied by the weights of features from the current split.\n\nNon-negative float values are supported for each weight.\n\nSupported formats for setting the value of this parameter:\n\n### [first\\_feature\\_use\\_penalties](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#first_feature_use_penalties)\nCommand-line: `--first-feature-use-penalties`\n\nPer-feature penalties for the first occurrence of the feature in the model. The given value is subtracted from the score if the current candidate is the first one to include the feature in the model.\n\nRefer to the [Per-object and per-feature penalties](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/algorithm-score-functions) section for details on applying different score penalties.\n\nNon-negative float values are supported for each penalty.\n\n### [fixed\\_binary\\_splits](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#fixed_binary_splits)\nCommand-line: `--fixed-binary-splits`\n\nA list of indices of binary features to put at the top of each tree.\n\n### [penalties\\_coefficient](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#penalties_coefficient)\nCommand-line: `--penalties-coefficient`\n\nA single-value common coefficient to multiply all penalties.\n\n### [per\\_object\\_feature\\_penalties](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#per_object_feature_penalties)\nCommand-line: `--per-object-feature-penalties`\n\nPer-object penalties for the first use of the feature for the object. The given value is multiplied by the number of objects that are divided by the current split and use the feature for the first time.\n\nRefer to the [Per-object and per-feature penalties](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/algorithm-score-functions) section for details on applying different score penalties.\n\nNon-negative float values are supported for each penalty.\n\n### [model\\_shrink\\_rate](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#model_shrink_rate)\nCommand-line: `--model-shrink-rate`\n\nThe constant used to calculate the coefficient for multiplying the model on each iteration.\n\n### [model\\_shrink\\_mode](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common#model_shrink_mode)\nCommand-line: `model_shrink_mode`\n\nDetermines how the actual model shrinkage coefficient is calculated at each iteration.\n\n## CTR settings\n### [simple\\_ctr](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/ctr#simple_ctr)\nQuantization settings for simple [categorical features](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/algorithm-main-stages_cat-to-numberic). Use this parameter to specify the principles for defining the class of the object for regression tasks. By default, it is considered that an object belongs to the positive class if its' label value is greater than the median of all label values of the dataset.\n\n### [combinations\\_ctr](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/ctr#combinations_ctr)\nQuantization settings for combinations of [categorical features](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/algorithm-main-stages_cat-to-numberic).\n\n### [per\\_feature\\_ctr](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/ctr#per_feature_ctr)\nPer-feature quantization settings for [categorical features](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/algorithm-main-stages_cat-to-numberic).\n\n### [ctr\\_target\\_border\\_count](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/ctr#ctr_target_border_count)\nThe maximum number of borders to use in target quantization for categorical features that need it. Allowed values are integers from 1 to 255 inclusively.\n\n### [counter\\_calc\\_method](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/ctr#counter_calc_method)\nThe method for calculating the Counter CTR type.\n\n### [max\\_ctr\\_complexity](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/ctr#max_ctr_complexity)\nThe maximum number of features that can be combined.\n\n### [ctr\\_leaf\\_count\\_limit](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/ctr#ctr_leaf_count_limit)\nThe maximum number of leaves with categorical features. If the quantity exceeds the specified value a part of leaves is discarded.\n\n### [store\\_all\\_simple\\_ctr](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/ctr#store_all_simple_ctr)\nIgnore categorical features, which are not used in feature combinations, when choosing candidates for exclusion.\n\n### [final\\_ctr\\_computation\\_mode](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/ctr#final_ctr_computation_mode)\nFinal CTR computation mode.\n\n## Input file settings\n### [\\-f, --learn-set](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/input#-f)\nThe path to the input file that contains the dataset description.\n\n### [\\-t, --test-set](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/input#-t)\nA comma-separated list of input files that contain the validation dataset description (the format must be the same as used in the training dataset).\n\n### [\\--cd, --column-description](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/input#--cd)\nThe path to the input file that contains the [columns description](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/input-data_column-descfile).\n\n### [\\--learn-pairs](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/input#--learn-pairs)\nThe path to the input file that contains the [pairs description](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/input-data_pairs-description) for the training dataset.\n\n### [\\--test-pairs](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/input#--test-pairs)\nThe path to the input file that contains the [pairs description](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/input-data_pairs-description) for the validation dataset.\n\n### [\\--learn-group-weights](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/input#--learn-group-weights)\nThe path to the input file that contains the weights of groups. Refer to the [Group weights](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/input-data_group-weights) section for format details.\n\n### [\\--test-group-weights](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/input#--test-group-weights)\nThe path to the input file that contains the weights of groups for the validation dataset.\n\n### [\\--learn-baseline](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/input#--learn-baseline)\nThe path to the input file that contains baseline values for the training dataset.\n\n### [\\--test-baseline](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/input#--test-baseline)\nThe path to the input file that contains baseline values for the validation dataset.\n\n### [\\--delimiter](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/input#--delimiter)\nThe delimiter character used to separate the data in the dataset description input file.\n\n### [\\--has-header](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/input#--has-header)\nRead the column names from the first line of the dataset description file if this parameter is set.\n\n### [\\--params-files](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/input#--params-files)\nThe path to the input JSON file that contains the training parameters, for example:\n\n### [\\--nan-mode](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/input#--nan-mode)\nThe method for [processing missing values](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/algorithm-missing-values-processing) in the input dataset.\n\n## Multiclassification settings\n### [classes\\_count](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/multiclassification#classes_count)\nCommand-line: `--classes-count`\n\nThe upper limit for the numeric class label. Defines the number of classes for multiclassification.\n\n### [\\--class-names](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/multiclassification#--class-names)\nThis parameter is only for Command-line.\n\nClasses names. Allows to redefine the default values when using the MultiClass and Logloss metrics.\n\n## Output settings\n### [logging\\_level](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/output#logging_level)\nCommand line: `--logging-level`\n\nThe logging level to output to stdout.\n\n### [metric\\_period](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/output#metric_period)\nCommand line: `--metric-period`\n\nThe frequency of iterations to calculate the values of [objectives and metrics](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/loss-functions).\n\nThe usage of this parameter speeds up the training.\n\n### [verbose](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/output#verbose)\nCommand line: `--verbose`\n\n*Alias:*`verbose_eval`\n\nThe purpose of this parameter depends on the type of the given value:\n\n### [train\\_dir](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/output#train_dir)\nCommand line: `--train-dir`\n\nThe directory for storing the files generated during training.\n\n### [model\\_size\\_reg](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/output#model_size_reg)\nCommand line: `--model-size-reg`\n\nThe model size regularization coefficient. The larger the value, the smaller the model size. Refer to the [Model size regularization coefficient](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/model-size-reg) section for details.\n\nThis regularization is needed only for models with categorical features (other models are small).\n\n### [allow\\_writing\\_files](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/output#allow_writing_files)\nAllow to write analytical and snapshot files during training.\n\n### [save\\_snapshot](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/output#save_snapshot)\nEnable snapshotting for [restoring the training progress after an interruption](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/features\/snapshots).\n\n### [snapshot\\_file](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/output#snapshot_file)\nThe name of the file to save the training progress information in. This file is used for [recovering training after an interruption](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/features\/snapshots).\n\n### [snapshot\\_interval](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/output#snapshot_interval)\nThe interval between saving snapshots in seconds.\n\n### [roc\\_file](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/output#roc_file)\nThe name of the [output file](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/output-data_roc-curve-points) to save the ROC curve points to.\n\n## Overfitting detection settings\n### [early\\_stopping\\_rounds](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/overfitting-detection#early_stopping_rounds)\nSets the overfitting detector type to Iter and stops the training after the specified number of iterations since the iteration with the optimal metric value.\n\n### [od\\_type](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/overfitting-detection#od_type)\nCommand-line: `--od-type`\n\nThe type of the overfitting detector to use.\n\n### [od\\_pval](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/overfitting-detection#od_pval)\nCommand-line: `--od-pval`\n\nThe threshold for the IncToDec [overfitting detector](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/overfitting-detector) type.\n\n### [od\\_wait](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/overfitting-detection#od_wait)\nCommand-line: `--od-wait`\n\nThe number of iterations to continue the training after the iteration with the optimal metric value.\n\n## Performance settings\n### [thread\\_count](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/performance#thread_count)\nCommand-line: `-T`, `--thread-count`\n\nThe number of threads to use during the training.\n\n### [used\\_ram\\_limit](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/performance#used_ram_limit)\nCommand-line: `--used-ram-limit`\n\nAttempt to limit the amount of used CPU RAM.\n\n### [gpu\\_ram\\_part](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/performance#gpu_ram_part)\nCommand-line: `--gpu-ram-part`\n\nHow much of the GPU RAM to use for training.\n\n### [pinned\\_memory\\_size](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/performance#pinned_memory_size)\nCommand-line: `--pinned-memory-size`\n\nHow much pinned (page-locked) CPU RAM to use per GPU.\n\n### [gpu\\_cat\\_features\\_storage](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/performance#gpu_cat_features_storage)\nCommand-line: `--gpu-cat-features-storage`\n\nThe method for storing the categorical features' values.\n\n### [data\\_partition](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/performance#data_partition)\nCommand-line: `--data-partition`\n\nThe method for splitting the input dataset between multiple workers.\n\n## Processing unit settings\n### [task\\_type](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/processing-unit#task_type)\nCommand line: `--task-type`\n\nThe processing unit type to use for training.\n\n### [devices](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/#devices)\nCommand line: `--devices`\n\nIDs of the GPU devices to use for training (indices are zero-based).\n\n## Quantization settings\n### [target\\_border](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/quantization#target_border)\nCommand-line: `--target-border`\n\nIf set, defines the border for converting target values to 0 and 1.\n\n### [border\\_count](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/quantization#border_count)\nCommand-line: `-x`, `--border-count`\n\n*Alias:* `max_bin`\n\nThe number of splits for numerical features. Allowed values are integers from 1 to 65535 inclusively.\n\n### [feature\\_border\\_type](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/quantization#feature_border_type)\nCommand-line: `--feature-border-type`\n\nThe [quantization mode](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/concepts\/quantization) for numerical features.\n\n### [per\\_float\\_feature\\_quantization](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/quantization#per_float_feature_quantization)\nCommand-line: `--per-float-feature-quantization`\n\nThe quantization description for the specified feature or list of features.\n\n## Text processing parameters\nThese parameters are only for the Python package and Command-line version.\n\n### [tokenizers](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/text-processing#tokenizers)\nCommand-line: `--tokenizers`\n\nTokenizers used to preprocess Text type feature columns before creating the dictionary.\n\n### [dictionaries](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/text-processing#dictionaries)\nCommand-line: `--dictionaries`\n\nDictionaries used to preprocess Text type feature columns.\n\nFormat:\n\n### [feature\\_calcers](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/text-processing#feature_calcers)\nCommand-line: `--feature-calcers`\n\nFeature calcers used to calculate new features based on preprocessed Text type feature columns.\n\nFormat:\n\n### [text\\_processing](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/text-processing#text_processing)\nCommand-line: `--text-processing`\n\nA JSON specification of tokenizers, dictionaries and feature calcers, which determine how text features are converted into a list of float features.\n\n[Example](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/text-processing__specification-example)\n\nRefer to the description of the following parameters for details on supported values:\n\n## Visualization settings\nThese parameters are only for the Python package.\n\n### [name](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/visualization#name)\nThe experiment name to display in [visualization tools](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/features\/visualization).\n\n### Was the article helpful?\nYes\n\nNo\n\nPrevious\n\n[Object importances](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/features\/object-importances-calcution)\n\nNext\n\n[Common parameters](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/en\/references\/training-parameters\/common)\n\n![](https:\/\/mc.yandex.ru\/watch\/60763294)","attrs_readable_markdown":"These parameters are for the Python package, R package and Command-line version.\n\nFor the Python package several parameters have aliases. For example, the `--iterations` parameter has the following synonyms: `num_boost_round`, `n_estimators`, `num_trees`. Simultaneous usage of different names of one parameter raises an error.\n\nTraining or inference on CUDA-enabled GPUs requires NVIDIA Driver of version 450.80.02 or higher.\n\n## Common parameters\n### [loss\\_function](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#loss_function)\nCommand-line: `--loss-function`\n\n*Alias:* `objective`\n\nThe [metric](https:\/\/catboost.ai\/docs\/en\/concepts\/loss-functions) to use in training. The specified value also determines the machine learning problem to solve. Some metrics support optional parameters (see the [Objectives and metrics](https:\/\/catboost.ai\/docs\/en\/concepts\/loss-functions) section for details on each metric).\n\n### [custom\\_metric](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#custom_metric)\nCommand-line: `--custom-metric`\n\n[Metric](https:\/\/catboost.ai\/docs\/en\/concepts\/loss-functions) values to output during training. These functions are not optimized and are displayed for informational purposes only. Some metrics support optional parameters (see the [Objectives and metrics](https:\/\/catboost.ai\/docs\/en\/concepts\/loss-functions) section for details on each metric).\n\n### [eval\\_metric](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#eval_metric)\nCommand-line: `--eval-metric`\n\nThe metric used for overfitting detection (if enabled) and best model selection (if enabled). Some metrics support optional parameters (see the [Objectives and metrics](https:\/\/catboost.ai\/docs\/en\/concepts\/loss-functions) section for details on each metric).\n\n### [iterations](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#iterations)\nCommand-line: `-i`, `--iterations`\n\n*Aliases:* `num_boost_round`, `n_estimators`, `num_trees`\n\nThe maximum number of trees that can be built when solving machine learning problems.\n\n### [learning\\_rate](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#learning_rate)\nCommand-line: `-w`, `--learning-rate`\n\n*Alias:* `eta`\n\nThe learning rate.\n\nUsed for reducing the gradient step.\n\n### [random\\_seed](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#random_seed)\nCommand-line: `-r`, `--random-seed`\n\n*Alias:*`random_state`\n\nThe random seed used for training.\n\n### [l2\\_leaf\\_reg](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#l2_leaf_reg)\nCommand-line: `--l2-leaf-reg`, `l2-leaf-regularizer`\n\n*Alias:* `reg_lambda`\n\nCoefficient at the L2 regularization term of the cost function.\n\n### [bootstrap\\_type](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#bootstrap_type)\nCommand-line: `--bootstrap-type`\n\n[Bootstrap type](https:\/\/catboost.ai\/docs\/en\/concepts\/algorithm-main-stages_bootstrap-options). Defines the method for sampling the weights of objects.\n\n### [bagging\\_temperature](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#bagging_temperature)\nCommand-line: `--bagging-temperature`\n\nDefines the settings of the Bayesian bootstrap. It is used by default in classification and regression modes.\n\n### [subsample](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#subsample)\nCommand-line: `--subsample`\n\nSample rate for bagging.\n\n### [sampling\\_frequency](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#sampling_frequency)\nCommand-line: `--sampling-frequency`\n\nFrequency to sample weights and objects when building trees.\n\n### [sampling\\_unit](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#sampling_unit)\nCommand-line: `--sampling-unit`\n\nThe sampling scheme.\n\n### [mvs\\_reg](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#mvs_reg)\nCommand-line: `--mvs-reg`\n\nAffects the weight of the denominator and can be used for balancing between the importance and Bernoulli sampling (setting it to 0 implies importance sampling and to ∞ \\\\infty - Bernoulli).\n\n### [random\\_strength](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#random_strength)\nCommand-line: `--random-strength`\n\nThe amount of randomness to use for scoring splits when the tree structure is selected. Use this parameter to avoid overfitting the model.\n\n### [use\\_best\\_model](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#use_best_model)\nCommand-line: `--use-best-model`\n\nUse the validation dataset to identify the iteration with the optimal value of the evaluation metric.  \n Save only trees up to this iteration (inclusive) in the resulting model.\n\n### [best\\_model\\_min\\_trees](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#best_model_min_trees)\nCommand-line: `--best-model-min-trees`\n\nThe minimal number of trees that the best model should have. If set, the output model contains at least the given number of trees even if the optimal value of the evaluation metric on the validation dataset is achieved with smaller number of trees.\n\n### [depth](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#depth)\nCommand-line: `-n`, `--depth`\n\n*Alias:* `max_depth`\n\nDepth of the trees.\n\n### [grow\\_policy](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#grow_policy)\nCommand-line: `--grow-policy`\n\nThe tree growing policy. Defines how to perform greedy tree construction.\n\n### [min\\_data\\_in\\_leaf](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#min_data_in_leaf)\nCommand-line: `--min-data-in-leaf`\n\n*Alias:* `min_child_samples`\n\nThe minimum number of training samples in a leaf. CatBoost does not search for new splits in leaves with samples count less than the specified value.\n\n### [max\\_leaves](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#max_leavescommon.md#)\nCommand-line: `--max-leaves`\n\n*Alias:*`num_leaves`\n\nThe maximum number of leafs in the resulting tree. Can be used only with the Lossguide growing policy.\n\n### [ignored\\_features](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#ignored_features)\nCommand-line: `-I`, `--ignore-features`\n\nFeature indices or names to exclude from the training. It is assumed that all passed values are feature names if at least one of the passed values can not be converted to a number or a range of numbers. Otherwise, it is assumed that all passed values are feature indices.\n\nSpecifics:\n\n- Non-negative indices that do not match any features are successfully ignored. For example, if five features are defined for the objects in the dataset and this parameter is set to \"42\", the corresponding non-existing feature is successfully ignored.\n- The identifier corresponds to the feature's index. Feature indices used in train and feature importance are numbered from 0 to `featureCount – 1`. If a file is used as [input data](https:\/\/catboost.ai\/docs\/en\/concepts\/input-data) then any non-feature column types are ignored when calculating these indices. For example, each row in the input file contains data in the following order: `cat feature<\\t>label value<\\t>num feature`. So for the row `rock<\\t>0<\\t>42`, the identifier for the \"rock\" feature is 0, and for the \"42\" feature it's 1.\n- The addition of a non-existing feature name raises an error.\n### [one\\_hot\\_max\\_size](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#one_hot_max_size)\nCommand-line: `--one-hot-max-size`\n\nUse one-hot encoding for all categorical features with a number of different values less than or equal to the given parameter value. Ctrs are not calculated for such features.\n\n### [has\\_time](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#has_time)\nCommand-line: `--has-time`\n\nUse the order of objects in the input data (do not perform random permutations during the [Transforming categorical features to numerical features](https:\/\/catboost.ai\/docs\/en\/concepts\/algorithm-main-stages_cat-to-numberic) and [Choosing the tree structure](https:\/\/catboost.ai\/docs\/en\/concepts\/algorithm-main-stages_choose-tree-structure) stages).\n\n### [rsm](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#rsm)\nCommand-line: `--rsm`\n\n*Alias:*`colsample_bylevel`\n\nRandom subspace method. The percentage of features to use at each split selection, when features are selected over again at random.\n\n### [nan\\_mode](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#nan_mode)\nCommand-line: `--nan-mode`\n\nThe method for [processing missing values](https:\/\/catboost.ai\/docs\/en\/concepts\/algorithm-missing-values-processing) in the input dataset.\n\n### [input\\_borders](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#input_borders)\nCommand-line: `--input-borders-file`\n\nLoad [Custom quantization borders and missing value modes](https:\/\/catboost.ai\/docs\/en\/concepts\/input-data_custom-borders) from a file (do not generate them).\n\n### [output\\_borders](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#output_borders)\nCommand-line: `--output-borders-file`\n\nSave quantization borders for the current dataset to a file.\n\n### [fold\\_permutation\\_block](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#fold_permutation_block)\nCommand-line: `--fold-permutation-block`\n\nObjects in the dataset are grouped in blocks before the random permutations. This parameter defines the size of the blocks.\n\n### [leaf\\_estimation\\_method](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#leaf_estimation_method)\nCommand-line: `--leaf-estimation-method`\n\nThe method used to calculate the values in leaves.\n\n### [leaf\\_estimation\\_iterations](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#leaf_estimation_iterations)\nCommand-line: `--leaf-estimation-iterations`\n\nThis parameter regulates how many steps are done in every tree when calculating leaf values.\n\n### [leaf\\_estimation\\_backtracking](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#leaf_estimation_backtracking)\nCommand-line: `--leaf-estimation-backtracking`\n\nWhen the value of the `leaf_estimation_iterations` parameter is greater than 1, CatBoost makes several gradient or newton steps when calculating the resulting leaf values of a tree.\n\n### [fold\\_len\\_multiplier](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#fold_len_multiplier)\nCommand-line: `--fold-len-multiplier`\n\nCoefficient for changing the length of folds.\n\n### [approx\\_on\\_full\\_history](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#approx_on_full_history)\nCommand-line:`--approx-on-full-history`\n\nThe principles for calculating the approximated values.\n\n### [class\\_weights](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#class_weights)\nCommand-line: `--class-weights`\n\nClass weights. The values are used as multipliers for the object weights. This parameter can be used for solving binary classification and multiclassification problems.\n\n### [class\\_names](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#class_names)\nClasses names. Allows to redefine the default values when using the MultiClass and Logloss metrics.\n\n### [auto\\_class\\_weights](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#auto_class_weights)\nCommand-line: `--auto-class-weights`\n\nAutomatically calculate class weights based either on the total weight or the total number of objects in each class. The values are used as multipliers for the object weights.\n\nSupported values:\n\n- None — All class weights are set to 1\n- Balanced:\n  C W k \\= m a x c \\= 1 K ( ∑ t i \\= c w i ) ∑ t i \\= k w i CW\\_k=\\\\displaystyle\\\\frac{max\\_{c=1}^K(\\\\sum\\_{t\\_{i}=c}{w\\_i})}{\\\\sum\\_{t\\_{i}=k}{w\\_{i}}}\n- SqrtBalanced:\n  C W k \\= m a x c \\= 1 K ( ∑ t i \\= c w i ) ∑ t i \\= k w i CW\\_k=\\\\sqrt{\\\\displaystyle\\\\frac{max\\_{c=1}^K(\\\\sum\\_{t\\_i=c}{w\\_i})}{\\\\sum\\_{t\\_i=k}{w\\_i}}}\n### [scale\\_pos\\_weight](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#scale_pos_weight)\nThe weight for class 1 in binary classification. The value is used as a multiplier for the weights of objects from class 1.\n\n### [boosting\\_type](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#boosting_type)\nCommand-line: `--boosting-type`\n\nBoosting scheme.\n\n### [boost\\_from\\_average](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#boost_from_average)\nCommand-line: `--boost-from-average`\n\nInitialize approximate values by best constant value for the specified loss function.\n\n### [langevin](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#langevin)\nCommand-line: `--langevin`\n\nEnables the Stochastic Gradient Langevin Boosting mode.\n\n### [diffusion\\_temperature](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#diffusion_temperature)\nCommand-line: `--diffusion-temperature`\n\nThe diffusion temperature of the Stochastic Gradient Langevin Boosting mode.\n\n### [posterior\\_sampling](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#posterior_sampling)\nCommand-line: `--posterior-sampling`\n\nIf this parameter is set several options are specified as follows and model parameters are checked to obtain uncertainty predictions with good theoretical properties.\n\n### [allow\\_const\\_label](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#allow_const_label)\nCommand-line: `--allow-const-label`\n\nUse it to train models with datasets that have equal label values for all objects.\n\n### [score\\_function](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#score_function)\nCommand-line: `--score-function`\n\nThe [score type](https:\/\/catboost.ai\/docs\/en\/concepts\/algorithm-score-functions) used to select the next split during the tree construction.\n\n### [monotone\\_constraints](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#monotone_constraints)\nCommand-line: `--monotone-constraints`\n\nImpose monotonic constraints on numerical features.\n\nPossible values:\n\n- \"1\" — Increasing constraint on the feature. The algorithm forces the model to be a non-decreasing function of this features.\n- \"\\-1\" — Decreasing constraint on the feature. The algorithm forces the model to be a non-increasing function of this features.\n- \"0\" — constraints are disabled.\n### [feature\\_weights](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#feature_weights)\nCommand-line: `--feature-weights`\n\nPer-feature multiplication weights used when choosing the best split. The score of each candidate is multiplied by the weights of features from the current split.\n\nNon-negative float values are supported for each weight.\n\nSupported formats for setting the value of this parameter:\n\n### [first\\_feature\\_use\\_penalties](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#first_feature_use_penalties)\nCommand-line: `--first-feature-use-penalties`\n\nPer-feature penalties for the first occurrence of the feature in the model. The given value is subtracted from the score if the current candidate is the first one to include the feature in the model.\n\nRefer to the [Per-object and per-feature penalties](https:\/\/catboost.ai\/docs\/en\/concepts\/algorithm-score-functions) section for details on applying different score penalties.\n\nNon-negative float values are supported for each penalty.\n\n### [fixed\\_binary\\_splits](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#fixed_binary_splits)\nCommand-line: `--fixed-binary-splits`\n\nA list of indices of binary features to put at the top of each tree.\n\n### [penalties\\_coefficient](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#penalties_coefficient)\nCommand-line: `--penalties-coefficient`\n\nA single-value common coefficient to multiply all penalties.\n\n### [per\\_object\\_feature\\_penalties](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#per_object_feature_penalties)\nCommand-line: `--per-object-feature-penalties`\n\nPer-object penalties for the first use of the feature for the object. The given value is multiplied by the number of objects that are divided by the current split and use the feature for the first time.\n\nRefer to the [Per-object and per-feature penalties](https:\/\/catboost.ai\/docs\/en\/concepts\/algorithm-score-functions) section for details on applying different score penalties.\n\nNon-negative float values are supported for each penalty.\n\n### [model\\_shrink\\_rate](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#model_shrink_rate)\nCommand-line: `--model-shrink-rate`\n\nThe constant used to calculate the coefficient for multiplying the model on each iteration.\n\n### [model\\_shrink\\_mode](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/common#model_shrink_mode)\nCommand-line: `model_shrink_mode`\n\nDetermines how the actual model shrinkage coefficient is calculated at each iteration.\n\n## CTR settings\n### [simple\\_ctr](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/ctr#simple_ctr)\nQuantization settings for simple [categorical features](https:\/\/catboost.ai\/docs\/en\/concepts\/algorithm-main-stages_cat-to-numberic). Use this parameter to specify the principles for defining the class of the object for regression tasks. By default, it is considered that an object belongs to the positive class if its' label value is greater than the median of all label values of the dataset.\n\n### [combinations\\_ctr](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/ctr#combinations_ctr)\nQuantization settings for combinations of [categorical features](https:\/\/catboost.ai\/docs\/en\/concepts\/algorithm-main-stages_cat-to-numberic).\n\n### [per\\_feature\\_ctr](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/ctr#per_feature_ctr)\nPer-feature quantization settings for [categorical features](https:\/\/catboost.ai\/docs\/en\/concepts\/algorithm-main-stages_cat-to-numberic).\n\n### [ctr\\_target\\_border\\_count](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/ctr#ctr_target_border_count)\nThe maximum number of borders to use in target quantization for categorical features that need it. Allowed values are integers from 1 to 255 inclusively.\n\n### [counter\\_calc\\_method](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/ctr#counter_calc_method)\nThe method for calculating the Counter CTR type.\n\n### [max\\_ctr\\_complexity](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/ctr#max_ctr_complexity)\nThe maximum number of features that can be combined.\n\n### [ctr\\_leaf\\_count\\_limit](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/ctr#ctr_leaf_count_limit)\nThe maximum number of leaves with categorical features. If the quantity exceeds the specified value a part of leaves is discarded.\n\n### [store\\_all\\_simple\\_ctr](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/ctr#store_all_simple_ctr)\nIgnore categorical features, which are not used in feature combinations, when choosing candidates for exclusion.\n\n### [final\\_ctr\\_computation\\_mode](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/ctr#final_ctr_computation_mode)\nFinal CTR computation mode.\n\n## Input file settings\n### [\\-f, --learn-set](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/input#-f)\nThe path to the input file that contains the dataset description.\n\n### [\\-t, --test-set](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/input#-t)\nA comma-separated list of input files that contain the validation dataset description (the format must be the same as used in the training dataset).\n\n### [\\--cd, --column-description](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/input#--cd)\nThe path to the input file that contains the [columns description](https:\/\/catboost.ai\/docs\/en\/concepts\/input-data_column-descfile).\n\n### [\\--learn-pairs](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/input#--learn-pairs)\nThe path to the input file that contains the [pairs description](https:\/\/catboost.ai\/docs\/en\/concepts\/input-data_pairs-description) for the training dataset.\n\n### [\\--test-pairs](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/input#--test-pairs)\nThe path to the input file that contains the [pairs description](https:\/\/catboost.ai\/docs\/en\/concepts\/input-data_pairs-description) for the validation dataset.\n\n### [\\--learn-group-weights](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/input#--learn-group-weights)\nThe path to the input file that contains the weights of groups. Refer to the [Group weights](https:\/\/catboost.ai\/docs\/en\/concepts\/input-data_group-weights) section for format details.\n\n### [\\--test-group-weights](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/input#--test-group-weights)\nThe path to the input file that contains the weights of groups for the validation dataset.\n\n### [\\--learn-baseline](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/input#--learn-baseline)\nThe path to the input file that contains baseline values for the training dataset.\n\n### [\\--test-baseline](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/input#--test-baseline)\nThe path to the input file that contains baseline values for the validation dataset.\n\n### [\\--delimiter](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/input#--delimiter)\nThe delimiter character used to separate the data in the dataset description input file.\n\nRead the column names from the first line of the dataset description file if this parameter is set.\n\n### [\\--params-files](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/input#--params-files)\nThe path to the input JSON file that contains the training parameters, for example:\n\n### [\\--nan-mode](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/input#--nan-mode)\nThe method for [processing missing values](https:\/\/catboost.ai\/docs\/en\/concepts\/algorithm-missing-values-processing) in the input dataset.\n\n## Multiclassification settings\n### [classes\\_count](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/multiclassification#classes_count)\nCommand-line: `--classes-count`\n\nThe upper limit for the numeric class label. Defines the number of classes for multiclassification.\n\n### [\\--class-names](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/multiclassification#--class-names)\nThis parameter is only for Command-line.\n\nClasses names. Allows to redefine the default values when using the MultiClass and Logloss metrics.\n\n## Output settings\n### [logging\\_level](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/output#logging_level)\nCommand line: `--logging-level`\n\nThe logging level to output to stdout.\n\n### [metric\\_period](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/output#metric_period)\nCommand line: `--metric-period`\n\nThe frequency of iterations to calculate the values of [objectives and metrics](https:\/\/catboost.ai\/docs\/en\/concepts\/loss-functions).\n\nThe usage of this parameter speeds up the training.\n\n### [verbose](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/output#verbose)\nCommand line: `--verbose`\n\n*Alias:*`verbose_eval`\n\nThe purpose of this parameter depends on the type of the given value:\n\n### [train\\_dir](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/output#train_dir)\nCommand line: `--train-dir`\n\nThe directory for storing the files generated during training.\n\n### [model\\_size\\_reg](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/output#model_size_reg)\nCommand line: `--model-size-reg`\n\nThe model size regularization coefficient. The larger the value, the smaller the model size. Refer to the [Model size regularization coefficient](https:\/\/catboost.ai\/docs\/en\/references\/model-size-reg) section for details.\n\nThis regularization is needed only for models with categorical features (other models are small).\n\n### [allow\\_writing\\_files](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/output#allow_writing_files)\nAllow to write analytical and snapshot files during training.\n\n### [save\\_snapshot](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/output#save_snapshot)\nEnable snapshotting for [restoring the training progress after an interruption](https:\/\/catboost.ai\/docs\/en\/features\/snapshots).\n\n### [snapshot\\_file](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/output#snapshot_file)\nThe name of the file to save the training progress information in. This file is used for [recovering training after an interruption](https:\/\/catboost.ai\/docs\/en\/features\/snapshots).\n\n### [snapshot\\_interval](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/output#snapshot_interval)\nThe interval between saving snapshots in seconds.\n\n### [roc\\_file](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/output#roc_file)\nThe name of the [output file](https:\/\/catboost.ai\/docs\/en\/concepts\/output-data_roc-curve-points) to save the ROC curve points to.\n\n## Overfitting detection settings\n### [early\\_stopping\\_rounds](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/overfitting-detection#early_stopping_rounds)\nSets the overfitting detector type to Iter and stops the training after the specified number of iterations since the iteration with the optimal metric value.\n\n### [od\\_type](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/overfitting-detection#od_type)\nCommand-line: `--od-type`\n\nThe type of the overfitting detector to use.\n\n### [od\\_pval](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/overfitting-detection#od_pval)\nCommand-line: `--od-pval`\n\nThe threshold for the IncToDec [overfitting detector](https:\/\/catboost.ai\/docs\/en\/concepts\/overfitting-detector) type.\n\n### [od\\_wait](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/overfitting-detection#od_wait)\nCommand-line: `--od-wait`\n\nThe number of iterations to continue the training after the iteration with the optimal metric value.\n\n## Performance settings\n### [thread\\_count](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/performance#thread_count)\nCommand-line: `-T`, `--thread-count`\n\nThe number of threads to use during the training.\n\n### [used\\_ram\\_limit](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/performance#used_ram_limit)\nCommand-line: `--used-ram-limit`\n\nAttempt to limit the amount of used CPU RAM.\n\n### [gpu\\_ram\\_part](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/performance#gpu_ram_part)\nCommand-line: `--gpu-ram-part`\n\nHow much of the GPU RAM to use for training.\n\n### [pinned\\_memory\\_size](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/performance#pinned_memory_size)\nCommand-line: `--pinned-memory-size`\n\nHow much pinned (page-locked) CPU RAM to use per GPU.\n\n### [gpu\\_cat\\_features\\_storage](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/performance#gpu_cat_features_storage)\nCommand-line: `--gpu-cat-features-storage`\n\nThe method for storing the categorical features' values.\n\n### [data\\_partition](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/performance#data_partition)\nCommand-line: `--data-partition`\n\nThe method for splitting the input dataset between multiple workers.\n\n## Processing unit settings\n### [task\\_type](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/processing-unit#task_type)\nCommand line: `--task-type`\n\nThe processing unit type to use for training.\n\n### [devices](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/#devices)\nCommand line: `--devices`\n\nIDs of the GPU devices to use for training (indices are zero-based).\n\n## Quantization settings\n### [target\\_border](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/quantization#target_border)\nCommand-line: `--target-border`\n\nIf set, defines the border for converting target values to 0 and 1.\n\n### [border\\_count](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/quantization#border_count)\nCommand-line: `-x`, `--border-count`\n\n*Alias:* `max_bin`\n\nThe number of splits for numerical features. Allowed values are integers from 1 to 65535 inclusively.\n\n### [feature\\_border\\_type](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/quantization#feature_border_type)\nCommand-line: `--feature-border-type`\n\nThe [quantization mode](https:\/\/catboost.ai\/docs\/en\/concepts\/quantization) for numerical features.\n\n### [per\\_float\\_feature\\_quantization](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/quantization#per_float_feature_quantization)\nCommand-line: `--per-float-feature-quantization`\n\nThe quantization description for the specified feature or list of features.\n\n## Text processing parameters\nThese parameters are only for the Python package and Command-line version.\n\n### [tokenizers](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/text-processing#tokenizers)\nCommand-line: `--tokenizers`\n\nTokenizers used to preprocess Text type feature columns before creating the dictionary.\n\n### [dictionaries](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/text-processing#dictionaries)\nCommand-line: `--dictionaries`\n\nDictionaries used to preprocess Text type feature columns.\n\nFormat:\n\n### [feature\\_calcers](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/text-processing#feature_calcers)\nCommand-line: `--feature-calcers`\n\nFeature calcers used to calculate new features based on preprocessed Text type feature columns.\n\nFormat:\n\n### [text\\_processing](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/text-processing#text_processing)\nCommand-line: `--text-processing`\n\nA JSON specification of tokenizers, dictionaries and feature calcers, which determine how text features are converted into a list of float features.\n\n[Example](https:\/\/catboost.ai\/docs\/en\/references\/text-processing__specification-example)\n\nRefer to the description of the following parameters for details on supported values:\n\n## Visualization settings\nThese parameters are only for the Python package.\n\n### [name](https:\/\/catboost.ai\/docs\/en\/references\/training-parameters\/visualization#name)\nThe experiment name to display in [visualization tools](https:\/\/catboost.ai\/docs\/en\/features\/visualization).","meta_canonical":null}

3. Robots.txt Check

Query:

Response:

4. Spam/Ban Check

Query:

Response:

5. Seen Status Check

ℹ️ Skipped - page is already crawled

📄

INDEXABLE

✅

CRAWLED

1 day ago

🤖

ROBOTS ALLOWED

Page Info Filters

Filter	Status	Condition	Details
HTTP status	PASS	`download_http_code = 200`	HTTP 200
Age cutoff	PASS	`download_stamp > now() - 6 MONTH`	0.1 months ago
History drop	PASS	`isNull(history_drop_reason)`	No drop reason
Spam/ban	PASS	`fh_dont_index != 1 AND ml_spam_score = 0`	ml_spam_score=0
Canonical	PASS	`meta_canonical IS NULL OR = '' OR = src_unparsed`	Not set

Page Details

Property	Value
URL	https://catboost.ai/docs/en/references/training-parameters/
Last Crawled	2026-04-12 12:44:21 (1 day ago)
First Indexed	2024-11-15 15:36:19 (1 year ago)
HTTP Status Code	200
Meta Title	Overview \| CatBoost
Meta Description	These parameters are for the Python package, R package and Command-line version.
Meta Canonical	null
Boilerpipe Text	These parameters are for the Python package, R package and Command-line version. For the Python package several parameters have aliases. For example, the --iterations parameter has the following synonyms: num_boost_round , n_estimators , num_trees . Simultaneous usage of different names of one parameter raises an error. Training or inference on CUDA-enabled GPUs requires NVIDIA Driver of version 450.80.02 or higher. Common parameters loss_function Command-line: --loss-function Alias: objective The metric to use in training. The specified value also determines the machine learning problem to solve. Some metrics support optional parameters (see the Objectives and metrics section for details on each metric). custom_metric Command-line: --custom-metric Metric values to output during training. These functions are not optimized and are displayed for informational purposes only. Some metrics support optional parameters (see the Objectives and metrics section for details on each metric). eval_metric Command-line: --eval-metric The metric used for overfitting detection (if enabled) and best model selection (if enabled). Some metrics support optional parameters (see the Objectives and metrics section for details on each metric). iterations Command-line: -i , --iterations Aliases: num_boost_round , n_estimators , num_trees The maximum number of trees that can be built when solving machine learning problems. learning_rate Command-line: -w , --learning-rate Alias: eta The learning rate. Used for reducing the gradient step. random_seed Command-line: -r , --random-seed Alias: random_state The random seed used for training. l2_leaf_reg Command-line: --l2-leaf-reg , l2-leaf-regularizer Alias: reg_lambda Coefficient at the L2 regularization term of the cost function. bootstrap_type Command-line: --bootstrap-type Bootstrap type . Defines the method for sampling the weights of objects. bagging_temperature Command-line: --bagging-temperature Defines the settings of the Bayesian bootstrap. It is used by default in classification and regression modes. subsample Command-line: --subsample Sample rate for bagging. sampling_frequency Command-line: --sampling-frequency Frequency to sample weights and objects when building trees. sampling_unit Command-line: --sampling-unit The sampling scheme. mvs_reg Command-line: --mvs-reg Affects the weight of the denominator and can be used for balancing between the importance and Bernoulli sampling (setting it to 0 implies importance sampling and to ∞ \infty - Bernoulli). random_strength Command-line: --random-strength The amount of randomness to use for scoring splits when the tree structure is selected. Use this parameter to avoid overfitting the model. use_best_model Command-line: --use-best-model Use the validation dataset to identify the iteration with the optimal value of the evaluation metric. Save only trees up to this iteration (inclusive) in the resulting model. best_model_min_trees Command-line: --best-model-min-trees The minimal number of trees that the best model should have. If set, the output model contains at least the given number of trees even if the optimal value of the evaluation metric on the validation dataset is achieved with smaller number of trees. depth Command-line: -n , --depth Alias: max_depth Depth of the trees. grow_policy Command-line: --grow-policy The tree growing policy. Defines how to perform greedy tree construction. min_data_in_leaf Command-line: --min-data-in-leaf Alias: min_child_samples The minimum number of training samples in a leaf. CatBoost does not search for new splits in leaves with samples count less than the specified value. max_leaves Command-line: --max-leaves Alias: num_leaves The maximum number of leafs in the resulting tree. Can be used only with the Lossguide growing policy. ignored_features Command-line: -I , --ignore-features Feature indices or names to exclude from the training. It is assumed that all passed values are feature names if at least one of the passed values can not be converted to a number or a range of numbers. Otherwise, it is assumed that all passed values are feature indices. Specifics: Non-negative indices that do not match any features are successfully ignored. For example, if five features are defined for the objects in the dataset and this parameter is set to 42 , the corresponding non-existing feature is successfully ignored. The identifier corresponds to the feature's index. Feature indices used in train and feature importance are numbered from 0 to featureCount – 1 . If a file is used as input data then any non-feature column types are ignored when calculating these indices. For example, each row in the input file contains data in the following order: cat feature<\t>label value<\t>num feature . So for the row rock<\t>0<\t>42 , the identifier for the rock feature is 0, and for the 42 feature it's 1. The addition of a non-existing feature name raises an error. one_hot_max_size Command-line: --one-hot-max-size Use one-hot encoding for all categorical features with a number of different values less than or equal to the given parameter value. Ctrs are not calculated for such features. has_time Command-line: --has-time Use the order of objects in the input data (do not perform random permutations during the Transforming categorical features to numerical features and Choosing the tree structure stages). rsm Command-line: --rsm Alias: colsample_bylevel Random subspace method. The percentage of features to use at each split selection, when features are selected over again at random. nan_mode Command-line: --nan-mode The method for processing missing values in the input dataset. input_borders Command-line: --input-borders-file Load Custom quantization borders and missing value modes from a file (do not generate them). output_borders Command-line: --output-borders-file Save quantization borders for the current dataset to a file. fold_permutation_block Command-line: --fold-permutation-block Objects in the dataset are grouped in blocks before the random permutations. This parameter defines the size of the blocks. leaf_estimation_method Command-line: --leaf-estimation-method The method used to calculate the values in leaves. leaf_estimation_iterations Command-line: --leaf-estimation-iterations This parameter regulates how many steps are done in every tree when calculating leaf values. leaf_estimation_backtracking Command-line: --leaf-estimation-backtracking When the value of the leaf_estimation_iterations parameter is greater than 1, CatBoost makes several gradient or newton steps when calculating the resulting leaf values of a tree. fold_len_multiplier Command-line: --fold-len-multiplier Coefficient for changing the length of folds. approx_on_full_history Command-line: --approx-on-full-history The principles for calculating the approximated values. class_weights Command-line: --class-weights Class weights. The values are used as multipliers for the object weights. This parameter can be used for solving binary classification and multiclassification problems. class_names Classes names. Allows to redefine the default values when using the MultiClass and Logloss metrics. auto_class_weights Command-line: --auto-class-weights Automatically calculate class weights based either on the total weight or the total number of objects in each class. The values are used as multipliers for the object weights. Supported values: None — All class weights are set to 1 Balanced: C W k = m a x c = 1 K ( ∑ t i = c w i ) ∑ t i = k w i CW_k=\displaystyle\frac{max_{c=1}^K(\sum_{t_{i}=c}{w_i})}{\sum_{t_{i}=k}{w_{i}}} SqrtBalanced: C W k = m a x c = 1 K ( ∑ t i = c w i ) ∑ t i = k w i CW_k=\sqrt{\displaystyle\frac{max_{c=1}^K(\sum_{t_i=c}{w_i})}{\sum_{t_i=k}{w_i}}} scale_pos_weight The weight for class 1 in binary classification. The value is used as a multiplier for the weights of objects from class 1. boosting_type Command-line: --boosting-type Boosting scheme. boost_from_average Command-line: --boost-from-average Initialize approximate values by best constant value for the specified loss function. langevin Command-line: --langevin Enables the Stochastic Gradient Langevin Boosting mode. diffusion_temperature Command-line: --diffusion-temperature The diffusion temperature of the Stochastic Gradient Langevin Boosting mode. posterior_sampling Command-line: --posterior-sampling If this parameter is set several options are specified as follows and model parameters are checked to obtain uncertainty predictions with good theoretical properties. allow_const_label Command-line: --allow-const-label Use it to train models with datasets that have equal label values for all objects. score_function Command-line: --score-function The score type used to select the next split during the tree construction. monotone_constraints Command-line: --monotone-constraints Impose monotonic constraints on numerical features. Possible values: 1 — Increasing constraint on the feature. The algorithm forces the model to be a non-decreasing function of this features. -1 — Decreasing constraint on the feature. The algorithm forces the model to be a non-increasing function of this features. 0 — constraints are disabled. feature_weights Command-line: --feature-weights Per-feature multiplication weights used when choosing the best split. The score of each candidate is multiplied by the weights of features from the current split. Non-negative float values are supported for each weight. Supported formats for setting the value of this parameter: first_feature_use_penalties Command-line: --first-feature-use-penalties Per-feature penalties for the first occurrence of the feature in the model. The given value is subtracted from the score if the current candidate is the first one to include the feature in the model. Refer to the Per-object and per-feature penalties section for details on applying different score penalties. Non-negative float values are supported for each penalty. fixed_binary_splits Command-line: --fixed-binary-splits A list of indices of binary features to put at the top of each tree. penalties_coefficient Command-line: --penalties-coefficient A single-value common coefficient to multiply all penalties. per_object_feature_penalties Command-line: --per-object-feature-penalties Per-object penalties for the first use of the feature for the object. The given value is multiplied by the number of objects that are divided by the current split and use the feature for the first time. Refer to the Per-object and per-feature penalties section for details on applying different score penalties. Non-negative float values are supported for each penalty. model_shrink_rate Command-line: --model-shrink-rate The constant used to calculate the coefficient for multiplying the model on each iteration. model_shrink_mode Command-line: model_shrink_mode Determines how the actual model shrinkage coefficient is calculated at each iteration. CTR settings simple_ctr Quantization settings for simple categorical features . Use this parameter to specify the principles for defining the class of the object for regression tasks. By default, it is considered that an object belongs to the positive class if its' label value is greater than the median of all label values of the dataset. combinations_ctr Quantization settings for combinations of categorical features . per_feature_ctr Per-feature quantization settings for categorical features . ctr_target_border_count The maximum number of borders to use in target quantization for categorical features that need it. Allowed values are integers from 1 to 255 inclusively. counter_calc_method The method for calculating the Counter CTR type. max_ctr_complexity The maximum number of features that can be combined. ctr_leaf_count_limit The maximum number of leaves with categorical features. If the quantity exceeds the specified value a part of leaves is discarded. store_all_simple_ctr Ignore categorical features, which are not used in feature combinations, when choosing candidates for exclusion. final_ctr_computation_mode Final CTR computation mode. Input file settings -f, --learn-set The path to the input file that contains the dataset description. -t, --test-set A comma-separated list of input files that contain the validation dataset description (the format must be the same as used in the training dataset). --cd, --column-description The path to the input file that contains the columns description . --learn-pairs The path to the input file that contains the pairs description for the training dataset. --test-pairs The path to the input file that contains the pairs description for the validation dataset. --learn-group-weights The path to the input file that contains the weights of groups. Refer to the Group weights section for format details. --test-group-weights The path to the input file that contains the weights of groups for the validation dataset. --learn-baseline The path to the input file that contains baseline values for the training dataset. --test-baseline The path to the input file that contains baseline values for the validation dataset. --delimiter The delimiter character used to separate the data in the dataset description input file. Read the column names from the first line of the dataset description file if this parameter is set. --params-files The path to the input JSON file that contains the training parameters, for example: --nan-mode The method for processing missing values in the input dataset. Multiclassification settings classes_count Command-line: --classes-count The upper limit for the numeric class label. Defines the number of classes for multiclassification. --class-names This parameter is only for Command-line. Classes names. Allows to redefine the default values when using the MultiClass and Logloss metrics. Output settings logging_level Command line: --logging-level The logging level to output to stdout. metric_period Command line: --metric-period The frequency of iterations to calculate the values of objectives and metrics . The usage of this parameter speeds up the training. verbose Command line: --verbose Alias: verbose_eval The purpose of this parameter depends on the type of the given value: train_dir Command line: --train-dir The directory for storing the files generated during training. model_size_reg Command line: --model-size-reg The model size regularization coefficient. The larger the value, the smaller the model size. Refer to the Model size regularization coefficient section for details. This regularization is needed only for models with categorical features (other models are small). allow_writing_files Allow to write analytical and snapshot files during training. save_snapshot Enable snapshotting for restoring the training progress after an interruption . snapshot_file The name of the file to save the training progress information in. This file is used for recovering training after an interruption . snapshot_interval The interval between saving snapshots in seconds. roc_file The name of the output file to save the ROC curve points to. Overfitting detection settings early_stopping_rounds Sets the overfitting detector type to Iter and stops the training after the specified number of iterations since the iteration with the optimal metric value. od_type Command-line: --od-type The type of the overfitting detector to use. od_pval Command-line: --od-pval The threshold for the IncToDec overfitting detector type. od_wait Command-line: --od-wait The number of iterations to continue the training after the iteration with the optimal metric value. Performance settings thread_count Command-line: -T , --thread-count The number of threads to use during the training. used_ram_limit Command-line: --used-ram-limit Attempt to limit the amount of used CPU RAM. gpu_ram_part Command-line: --gpu-ram-part How much of the GPU RAM to use for training. pinned_memory_size Command-line: --pinned-memory-size How much pinned (page-locked) CPU RAM to use per GPU. gpu_cat_features_storage Command-line: --gpu-cat-features-storage The method for storing the categorical features' values. data_partition Command-line: --data-partition The method for splitting the input dataset between multiple workers. Processing unit settings task_type Command line: --task-type The processing unit type to use for training. devices Command line: --devices IDs of the GPU devices to use for training (indices are zero-based). Quantization settings target_border Command-line: --target-border If set, defines the border for converting target values to 0 and 1. border_count Command-line: -x , --border-count Alias: max_bin The number of splits for numerical features. Allowed values are integers from 1 to 65535 inclusively. feature_border_type Command-line: --feature-border-type The quantization mode for numerical features. per_float_feature_quantization Command-line: --per-float-feature-quantization The quantization description for the specified feature or list of features. Text processing parameters These parameters are only for the Python package and Command-line version. tokenizers Command-line: --tokenizers Tokenizers used to preprocess Text type feature columns before creating the dictionary. dictionaries Command-line: --dictionaries Dictionaries used to preprocess Text type feature columns. Format: feature_calcers Command-line: --feature-calcers Feature calcers used to calculate new features based on preprocessed Text type feature columns. Format: text_processing Command-line: --text-processing A JSON specification of tokenizers, dictionaries and feature calcers, which determine how text features are converted into a list of float features. Example Refer to the description of the following parameters for details on supported values: Visualization settings These parameters are only for the Python package. name The experiment name to display in visualization tools .
Markdown	[![Logo icon](https://yastatic.net/s3/locdoc/daas-static/catboost/71b237a322eec6f2889af0dae2a9c549.svg)](https://catboost.ai/ "CatBoost") - Installation - [Overview](https://catboost.ai/docs/en/references/training-parameters/en/concepts/installation) - Python package installation - CatBoost for Apache Spark installation - R package installation - Command-line version binary - Build from source - Key Features - Training parameters - [Overview](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/) - [Common parameters](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common) - [CTR settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/ctr) - [Input settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/input) - [Multiclassification settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/multiclassification) - [Output settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/output) - [Overfitting detection settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/overfitting-detection) - [Performance settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/performance) - [Processing unit settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/processing-unit) - [Quantization settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/quantization) - [Text processing parameters](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/text-processing) - [Visualization settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/visualization) - Python package - CatBoost for Apache Spark - R package - Command-line version - Applying models - Objectives and metrics - Model analysis - Data format description - [Parameter tuning](https://catboost.ai/docs/en/references/training-parameters/en/concepts/parameter-tuning) - [Speeding up the training](https://catboost.ai/docs/en/references/training-parameters/en/concepts/speed-up-training) - Data visualization - Algorithm details - [FAQ](https://catboost.ai/docs/en/references/training-parameters/en/concepts/faq) - Educational materials - [Development and contributions](https://catboost.ai/docs/en/references/training-parameters/en/concepts/development-and-contributions) - [Contacts](https://catboost.ai/docs/en/references/training-parameters/en/concepts/contacts) Overview ## In this article: - [Common parameters](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#common-parameters) - [loss\_function](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#loss_function) - [custom\_metric](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#custom_metric) - [eval\_metric](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#eval_metric) - [iterations](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#iterations) - [learning\_rate](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#learning_rate) - [random\_seed](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#random_seed) - [l2\_leaf\_reg](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#l2_leaf_reg) - [bootstrap\_type](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#bootstrap_type) - [bagging\_temperature](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#bagging_temperature) - [subsample](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#subsample) - [sampling\_frequency](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#sampling_frequency) - [sampling\_unit](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#sampling_unit) - [mvs\_reg](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#mvs_reg) - [random\_strength](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#random_strength) - [use\_best\_model](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#use_best_model) - [best\_model\_min\_trees](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#best_model_min_trees) - [depth](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#depth) - [grow\_policy](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#grow_policy) - [min\_data\_in\_leaf](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#min_data_in_leaf) - [max\_leaves](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#max_leaves) - [ignored\_features](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#ignored_features) - [one\_hot\_max\_size](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#one_hot_max_size) - [has\_time](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#has_time) - [rsm](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#rsm) - [nan\_mode](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#nan_mode) - [input\_borders](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#input_borders) - [output\_borders](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#output_borders) - [fold\_permutation\_block](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#fold_permutation_block) - [leaf\_estimation\_method](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#leaf_estimation_method) - [leaf\_estimation\_iterations](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#leaf_estimation_iterations) - [leaf\_estimation\_backtracking](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#leaf_estimation_backtracking) - [fold\_len\_multiplier](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#fold_len_multiplier) - [approx\_on\_full\_history](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#approx_on_full_history) - [class\_weights](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#class_weights) - [class\_names](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#class_names) - [auto\_class\_weights](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#auto_class_weights) - [scale\_pos\_weight](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#scale_pos_weight) - [boosting\_type](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#boosting_type) - [boost\_from\_average](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#boost_from_average) - [langevin](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#langevin) - [diffusion\_temperature](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#diffusion_temperature) - [posterior\_sampling](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#posterior_sampling) - [allow\_const\_label](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#allow_const_label) - [score\_function](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#score_function) - [monotone\_constraints](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#monotone_constraints) - [feature\_weights](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#feature_weights) - [first\_feature\_use\_penalties](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#first_feature_use_penalties) - [fixed\_binary\_splits](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#fixed_binary_splits) - [penalties\_coefficient](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#penalties_coefficient) - [per\_object\_feature\_penalties](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#per_object_feature_penalties) - [model\_shrink\_rate](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#model_shrink_rate) - [model\_shrink\_mode](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#model_shrink_mode) - [CTR settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#ctr-settings) - [simple\_ctr](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#simple_ctr) - [combinations\_ctr](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#combinations_ctr) - [per\_feature\_ctr](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#per_feature_ctr) - [ctr\_target\_border\_count](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#ctr_target_border_count) - [counter\_calc\_method](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#counter_calc_method) - [max\_ctr\_complexity](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#max_ctr_complexity) - [ctr\_leaf\_count\_limit](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#ctr_leaf_count_limit) - [store\_all\_simple\_ctr](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#store_all_simple_ctr) - [final\_ctr\_computation\_mode](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#final_ctr_computation_mode) - [Input file settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#input-file-settings) - [\-f, --learn-set](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#f,-learn-set) - [\-t, --test-set](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#t,-test-set) - [\--cd, --column-description](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#cd,-column-description) - [\--learn-pairs](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#learn-pairs) - [\--test-pairs](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#test-pairs) - [\--learn-group-weights](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#learn-group-weights) - [\--test-group-weights](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#test-group-weights) - [\--learn-baseline](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#learn-baseline) - [\--test-baseline](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#test-baseline) - [\--delimiter](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#delimiter) - [\--has-header](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#has-header) - [\--params-files](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#params-files) - [\--nan-mode](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#nan-mode) - [Multiclassification settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#multiclassification-settings) - [classes\_count](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#classes_count) - [\--class-names](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#class-names) - [Output settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#output-settings) - [logging\_level](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#logging_level) - [metric\_period](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#metric_period) - [verbose](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#verbose) - [train\_dir](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#train_dir) - [model\_size\_reg](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#model_size_reg) - [allow\_writing\_files](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#allow_writing_files) - [save\_snapshot](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#save_snapshot) - [snapshot\_file](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#snapshot_file) - [snapshot\_interval](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#snapshot_interval) - [roc\_file](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#roc_file) - [Overfitting detection settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#overfitting-detection-settings) - [early\_stopping\_rounds](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#early_stopping_rounds) - [od\_type](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#od_type) - [od\_pval](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#od_pval) - [od\_wait](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#od_wait) - [Performance settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#performance-settings) - [thread\_count](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#thread_count) - [used\_ram\_limit](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#used_ram_limit) - [gpu\_ram\_part](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#gpu_ram_part) - [pinned\_memory\_size](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#pinned_memory_size) - [gpu\_cat\_features\_storage](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#gpu_cat_features_storage) - [data\_partition](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#data_partition) - [Processing unit settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#processing-unit-settings) - [task\_type](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#task_type) - [devices](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#devices) - [Quantization settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#quantization-settings) - [target\_border](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#target_border) - [border\_count](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#border_count) - [feature\_border\_type](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#feature_border_type) - [per\_float\_feature\_quantization](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#per_float_feature_quantization) - [Text processing parameters](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#text-processing-parameters) - [tokenizers](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#tokenizers) - [dictionaries](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#dictionaries) - [feature\_calcers](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#feature_calcers) - [text\_processing](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#text_processing) - [Visualization settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#visualization-settings) - [name](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#name) 1. Training parameters 2. Overview # Overview - [Common parameters](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#common-parameters) - [loss\_function](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#loss_function) - [custom\_metric](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#custom_metric) - [eval\_metric](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#eval_metric) - [iterations](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#iterations) - [learning\_rate](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#learning_rate) - [random\_seed](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#random_seed) - [l2\_leaf\_reg](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#l2_leaf_reg) - [bootstrap\_type](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#bootstrap_type) - [bagging\_temperature](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#bagging_temperature) - [subsample](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#subsample) - [sampling\_frequency](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#sampling_frequency) - [sampling\_unit](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#sampling_unit) - [mvs\_reg](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#mvs_reg) - [random\_strength](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#random_strength) - [use\_best\_model](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#use_best_model) - [best\_model\_min\_trees](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#best_model_min_trees) - [depth](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#depth) - [grow\_policy](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#grow_policy) - [min\_data\_in\_leaf](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#min_data_in_leaf) - [max\_leaves](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#max_leaves) - [ignored\_features](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#ignored_features) - [one\_hot\_max\_size](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#one_hot_max_size) - [has\_time](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#has_time) - [rsm](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#rsm) - [nan\_mode](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#nan_mode) - [input\_borders](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#input_borders) - [output\_borders](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#output_borders) - [fold\_permutation\_block](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#fold_permutation_block) - [leaf\_estimation\_method](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#leaf_estimation_method) - [leaf\_estimation\_iterations](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#leaf_estimation_iterations) - [leaf\_estimation\_backtracking](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#leaf_estimation_backtracking) - [fold\_len\_multiplier](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#fold_len_multiplier) - [approx\_on\_full\_history](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#approx_on_full_history) - [class\_weights](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#class_weights) - [class\_names](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#class_names) - [auto\_class\_weights](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#auto_class_weights) - [scale\_pos\_weight](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#scale_pos_weight) - [boosting\_type](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#boosting_type) - [boost\_from\_average](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#boost_from_average) - [langevin](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#langevin) - [diffusion\_temperature](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#diffusion_temperature) - [posterior\_sampling](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#posterior_sampling) - [allow\_const\_label](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#allow_const_label) - [score\_function](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#score_function) - [monotone\_constraints](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#monotone_constraints) - [feature\_weights](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#feature_weights) - [first\_feature\_use\_penalties](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#first_feature_use_penalties) - [fixed\_binary\_splits](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#fixed_binary_splits) - [penalties\_coefficient](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#penalties_coefficient) - [per\_object\_feature\_penalties](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#per_object_feature_penalties) - [model\_shrink\_rate](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#model_shrink_rate) - [model\_shrink\_mode](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#model_shrink_mode) - [CTR settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#ctr-settings) - [simple\_ctr](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#simple_ctr) - [combinations\_ctr](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#combinations_ctr) - [per\_feature\_ctr](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#per_feature_ctr) - [ctr\_target\_border\_count](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#ctr_target_border_count) - [counter\_calc\_method](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#counter_calc_method) - [max\_ctr\_complexity](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#max_ctr_complexity) - [ctr\_leaf\_count\_limit](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#ctr_leaf_count_limit) - [store\_all\_simple\_ctr](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#store_all_simple_ctr) - [final\_ctr\_computation\_mode](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#final_ctr_computation_mode) - [Input file settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#input-file-settings) - [\-f, --learn-set](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#f,-learn-set) - [\-t, --test-set](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#t,-test-set) - [\--cd, --column-description](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#cd,-column-description) - [\--learn-pairs](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#learn-pairs) - [\--test-pairs](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#test-pairs) - [\--learn-group-weights](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#learn-group-weights) - [\--test-group-weights](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#test-group-weights) - [\--learn-baseline](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#learn-baseline) - [\--test-baseline](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#test-baseline) - [\--delimiter](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#delimiter) - [\--has-header](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#has-header) - [\--params-files](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#params-files) - [\--nan-mode](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#nan-mode) - [Multiclassification settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#multiclassification-settings) - [classes\_count](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#classes_count) - [\--class-names](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#class-names) - [Output settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#output-settings) - [logging\_level](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#logging_level) - [metric\_period](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#metric_period) - [verbose](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#verbose) - [train\_dir](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#train_dir) - [model\_size\_reg](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#model_size_reg) - [allow\_writing\_files](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#allow_writing_files) - [save\_snapshot](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#save_snapshot) - [snapshot\_file](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#snapshot_file) - [snapshot\_interval](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#snapshot_interval) - [roc\_file](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#roc_file) - [Overfitting detection settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#overfitting-detection-settings) - [early\_stopping\_rounds](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#early_stopping_rounds) - [od\_type](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#od_type) - [od\_pval](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#od_pval) - [od\_wait](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#od_wait) - [Performance settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#performance-settings) - [thread\_count](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#thread_count) - [used\_ram\_limit](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#used_ram_limit) - [gpu\_ram\_part](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#gpu_ram_part) - [pinned\_memory\_size](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#pinned_memory_size) - [gpu\_cat\_features\_storage](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#gpu_cat_features_storage) - [data\_partition](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#data_partition) - [Processing unit settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#processing-unit-settings) - [task\_type](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#task_type) - [devices](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#devices) - [Quantization settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#quantization-settings) - [target\_border](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#target_border) - [border\_count](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#border_count) - [feature\_border\_type](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#feature_border_type) - [per\_float\_feature\_quantization](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#per_float_feature_quantization) - [Text processing parameters](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#text-processing-parameters) - [tokenizers](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#tokenizers) - [dictionaries](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#dictionaries) - [feature\_calcers](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#feature_calcers) - [text\_processing](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#text_processing) - [Visualization settings](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#visualization-settings) - [name](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#name) These parameters are for the Python package, R package and Command-line version. For the Python package several parameters have aliases. For example, the `--iterations` parameter has the following synonyms: `num_boost_round`, `n_estimators`, `num_trees`. Simultaneous usage of different names of one parameter raises an error. Training or inference on CUDA-enabled GPUs requires NVIDIA Driver of version 450.80.02 or higher. ## Common parameters ### [loss\_function](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#loss_function) Command-line: `--loss-function` Alias: `objective` The [metric](https://catboost.ai/docs/en/references/training-parameters/en/concepts/loss-functions) to use in training. The specified value also determines the machine learning problem to solve. Some metrics support optional parameters (see the [Objectives and metrics](https://catboost.ai/docs/en/references/training-parameters/en/concepts/loss-functions) section for details on each metric). ### [custom\_metric](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#custom_metric) Command-line: `--custom-metric` [Metric](https://catboost.ai/docs/en/references/training-parameters/en/concepts/loss-functions) values to output during training. These functions are not optimized and are displayed for informational purposes only. Some metrics support optional parameters (see the [Objectives and metrics](https://catboost.ai/docs/en/references/training-parameters/en/concepts/loss-functions) section for details on each metric). ### [eval\_metric](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#eval_metric) Command-line: `--eval-metric` The metric used for overfitting detection (if enabled) and best model selection (if enabled). Some metrics support optional parameters (see the [Objectives and metrics](https://catboost.ai/docs/en/references/training-parameters/en/concepts/loss-functions) section for details on each metric). ### [iterations](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#iterations) Command-line: `-i`, `--iterations` Aliases: `num_boost_round`, `n_estimators`, `num_trees` The maximum number of trees that can be built when solving machine learning problems. ### [learning\_rate](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#learning_rate) Command-line: `-w`, `--learning-rate` Alias: `eta` The learning rate. Used for reducing the gradient step. ### [random\_seed](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#random_seed) Command-line: `-r`, `--random-seed` Alias:`random_state` The random seed used for training. ### [l2\_leaf\_reg](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#l2_leaf_reg) Command-line: `--l2-leaf-reg`, `l2-leaf-regularizer` Alias: `reg_lambda` Coefficient at the L2 regularization term of the cost function. ### [bootstrap\_type](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#bootstrap_type) Command-line: `--bootstrap-type` [Bootstrap type](https://catboost.ai/docs/en/references/training-parameters/en/concepts/algorithm-main-stages_bootstrap-options). Defines the method for sampling the weights of objects. ### [bagging\_temperature](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#bagging_temperature) Command-line: `--bagging-temperature` Defines the settings of the Bayesian bootstrap. It is used by default in classification and regression modes. ### [subsample](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#subsample) Command-line: `--subsample` Sample rate for bagging. ### [sampling\_frequency](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#sampling_frequency) Command-line: `--sampling-frequency` Frequency to sample weights and objects when building trees. ### [sampling\_unit](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#sampling_unit) Command-line: `--sampling-unit` The sampling scheme. ### [mvs\_reg](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#mvs_reg) Command-line: `--mvs-reg` Affects the weight of the denominator and can be used for balancing between the importance and Bernoulli sampling (setting it to 0 implies importance sampling and to ∞ \\infty ∞ - Bernoulli). ### [random\_strength](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#random_strength) Command-line: `--random-strength` The amount of randomness to use for scoring splits when the tree structure is selected. Use this parameter to avoid overfitting the model. ### [use\_best\_model](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#use_best_model) Command-line: `--use-best-model` Use the validation dataset to identify the iteration with the optimal value of the evaluation metric. Save only trees up to this iteration (inclusive) in the resulting model. ### [best\_model\_min\_trees](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#best_model_min_trees) Command-line: `--best-model-min-trees` The minimal number of trees that the best model should have. If set, the output model contains at least the given number of trees even if the optimal value of the evaluation metric on the validation dataset is achieved with smaller number of trees. ### [depth](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#depth) Command-line: `-n`, `--depth` Alias: `max_depth` Depth of the trees. ### [grow\_policy](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#grow_policy) Command-line: `--grow-policy` The tree growing policy. Defines how to perform greedy tree construction. ### [min\_data\_in\_leaf](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#min_data_in_leaf) Command-line: `--min-data-in-leaf` Alias: `min_child_samples` The minimum number of training samples in a leaf. CatBoost does not search for new splits in leaves with samples count less than the specified value. ### [max\_leaves](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#max_leavescommon.md#) Command-line: `--max-leaves` Alias:`num_leaves` The maximum number of leafs in the resulting tree. Can be used only with the Lossguide growing policy. ### [ignored\_features](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#ignored_features) Command-line: `-I`, `--ignore-features` Feature indices or names to exclude from the training. It is assumed that all passed values are feature names if at least one of the passed values can not be converted to a number or a range of numbers. Otherwise, it is assumed that all passed values are feature indices. Specifics: - Non-negative indices that do not match any features are successfully ignored. For example, if five features are defined for the objects in the dataset and this parameter is set to "42", the corresponding non-existing feature is successfully ignored. - The identifier corresponds to the feature's index. Feature indices used in train and feature importance are numbered from 0 to `featureCount – 1`. If a file is used as [input data](https://catboost.ai/docs/en/references/training-parameters/en/concepts/input-data) then any non-feature column types are ignored when calculating these indices. For example, each row in the input file contains data in the following order: `cat feature<\t>label value<\t>num feature`. So for the row `rock<\t>0<\t>42`, the identifier for the "rock" feature is 0, and for the "42" feature it's 1. - The addition of a non-existing feature name raises an error. ### [one\_hot\_max\_size](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#one_hot_max_size) Command-line: `--one-hot-max-size` Use one-hot encoding for all categorical features with a number of different values less than or equal to the given parameter value. Ctrs are not calculated for such features. ### [has\_time](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#has_time) Command-line: `--has-time` Use the order of objects in the input data (do not perform random permutations during the [Transforming categorical features to numerical features](https://catboost.ai/docs/en/references/training-parameters/en/concepts/algorithm-main-stages_cat-to-numberic) and [Choosing the tree structure](https://catboost.ai/docs/en/references/training-parameters/en/concepts/algorithm-main-stages_choose-tree-structure) stages). ### [rsm](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#rsm) Command-line: `--rsm` Alias:`colsample_bylevel` Random subspace method. The percentage of features to use at each split selection, when features are selected over again at random. ### [nan\_mode](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#nan_mode) Command-line: `--nan-mode` The method for [processing missing values](https://catboost.ai/docs/en/references/training-parameters/en/concepts/algorithm-missing-values-processing) in the input dataset. ### [input\_borders](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#input_borders) Command-line: `--input-borders-file` Load [Custom quantization borders and missing value modes](https://catboost.ai/docs/en/references/training-parameters/en/concepts/input-data_custom-borders) from a file (do not generate them). ### [output\_borders](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#output_borders) Command-line: `--output-borders-file` Save quantization borders for the current dataset to a file. ### [fold\_permutation\_block](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#fold_permutation_block) Command-line: `--fold-permutation-block` Objects in the dataset are grouped in blocks before the random permutations. This parameter defines the size of the blocks. ### [leaf\_estimation\_method](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#leaf_estimation_method) Command-line: `--leaf-estimation-method` The method used to calculate the values in leaves. ### [leaf\_estimation\_iterations](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#leaf_estimation_iterations) Command-line: `--leaf-estimation-iterations` This parameter regulates how many steps are done in every tree when calculating leaf values. ### [leaf\_estimation\_backtracking](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#leaf_estimation_backtracking) Command-line: `--leaf-estimation-backtracking` When the value of the `leaf_estimation_iterations` parameter is greater than 1, CatBoost makes several gradient or newton steps when calculating the resulting leaf values of a tree. ### [fold\_len\_multiplier](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#fold_len_multiplier) Command-line: `--fold-len-multiplier` Coefficient for changing the length of folds. ### [approx\_on\_full\_history](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#approx_on_full_history) Command-line:`--approx-on-full-history` The principles for calculating the approximated values. ### [class\_weights](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#class_weights) Command-line: `--class-weights` Class weights. The values are used as multipliers for the object weights. This parameter can be used for solving binary classification and multiclassification problems. ### [class\_names](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#class_names) Classes names. Allows to redefine the default values when using the MultiClass and Logloss metrics. ### [auto\_class\_weights](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#auto_class_weights) Command-line: `--auto-class-weights` Automatically calculate class weights based either on the total weight or the total number of objects in each class. The values are used as multipliers for the object weights. Supported values: - None — All class weights are set to 1 - Balanced: C W k \= m a x c \= 1 K ( ∑ t i \= c w i ) ∑ t i \= k w i CW\_k=\\displaystyle\\frac{max\_{c=1}^K(\\sum\_{t\_{i}=c}{w\_i})}{\\sum\_{t\_{i}=k}{w\_{i}}} CWk\=∑ti\=kwimaxc\=1K(∑ti\=cwi) - SqrtBalanced: C W k \= m a x c \= 1 K ( ∑ t i \= c w i ) ∑ t i \= k w i CW\_k=\\sqrt{\\displaystyle\\frac{max\_{c=1}^K(\\sum\_{t\_i=c}{w\_i})}{\\sum\_{t\_i=k}{w\_i}}} CWk\= ∑ti\=kwimaxc\=1K(∑ti\=cwi) ### [scale\_pos\_weight](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#scale_pos_weight) The weight for class 1 in binary classification. The value is used as a multiplier for the weights of objects from class 1. ### [boosting\_type](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#boosting_type) Command-line: `--boosting-type` Boosting scheme. ### [boost\_from\_average](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#boost_from_average) Command-line: `--boost-from-average` Initialize approximate values by best constant value for the specified loss function. ### [langevin](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#langevin) Command-line: `--langevin` Enables the Stochastic Gradient Langevin Boosting mode. ### [diffusion\_temperature](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#diffusion_temperature) Command-line: `--diffusion-temperature` The diffusion temperature of the Stochastic Gradient Langevin Boosting mode. ### [posterior\_sampling](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#posterior_sampling) Command-line: `--posterior-sampling` If this parameter is set several options are specified as follows and model parameters are checked to obtain uncertainty predictions with good theoretical properties. ### [allow\_const\_label](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#allow_const_label) Command-line: `--allow-const-label` Use it to train models with datasets that have equal label values for all objects. ### [score\_function](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#score_function) Command-line: `--score-function` The [score type](https://catboost.ai/docs/en/references/training-parameters/en/concepts/algorithm-score-functions) used to select the next split during the tree construction. ### [monotone\_constraints](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#monotone_constraints) Command-line: `--monotone-constraints` Impose monotonic constraints on numerical features. Possible values: - "1" — Increasing constraint on the feature. The algorithm forces the model to be a non-decreasing function of this features. - "\-1" — Decreasing constraint on the feature. The algorithm forces the model to be a non-increasing function of this features. - "0" — constraints are disabled. ### [feature\_weights](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#feature_weights) Command-line: `--feature-weights` Per-feature multiplication weights used when choosing the best split. The score of each candidate is multiplied by the weights of features from the current split. Non-negative float values are supported for each weight. Supported formats for setting the value of this parameter: ### [first\_feature\_use\_penalties](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#first_feature_use_penalties) Command-line: `--first-feature-use-penalties` Per-feature penalties for the first occurrence of the feature in the model. The given value is subtracted from the score if the current candidate is the first one to include the feature in the model. Refer to the [Per-object and per-feature penalties](https://catboost.ai/docs/en/references/training-parameters/en/concepts/algorithm-score-functions) section for details on applying different score penalties. Non-negative float values are supported for each penalty. ### [fixed\_binary\_splits](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#fixed_binary_splits) Command-line: `--fixed-binary-splits` A list of indices of binary features to put at the top of each tree. ### [penalties\_coefficient](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#penalties_coefficient) Command-line: `--penalties-coefficient` A single-value common coefficient to multiply all penalties. ### [per\_object\_feature\_penalties](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#per_object_feature_penalties) Command-line: `--per-object-feature-penalties` Per-object penalties for the first use of the feature for the object. The given value is multiplied by the number of objects that are divided by the current split and use the feature for the first time. Refer to the [Per-object and per-feature penalties](https://catboost.ai/docs/en/references/training-parameters/en/concepts/algorithm-score-functions) section for details on applying different score penalties. Non-negative float values are supported for each penalty. ### [model\_shrink\_rate](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#model_shrink_rate) Command-line: `--model-shrink-rate` The constant used to calculate the coefficient for multiplying the model on each iteration. ### [model\_shrink\_mode](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common#model_shrink_mode) Command-line: `model_shrink_mode` Determines how the actual model shrinkage coefficient is calculated at each iteration. ## CTR settings ### [simple\_ctr](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/ctr#simple_ctr) Quantization settings for simple [categorical features](https://catboost.ai/docs/en/references/training-parameters/en/concepts/algorithm-main-stages_cat-to-numberic). Use this parameter to specify the principles for defining the class of the object for regression tasks. By default, it is considered that an object belongs to the positive class if its' label value is greater than the median of all label values of the dataset. ### [combinations\_ctr](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/ctr#combinations_ctr) Quantization settings for combinations of [categorical features](https://catboost.ai/docs/en/references/training-parameters/en/concepts/algorithm-main-stages_cat-to-numberic). ### [per\_feature\_ctr](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/ctr#per_feature_ctr) Per-feature quantization settings for [categorical features](https://catboost.ai/docs/en/references/training-parameters/en/concepts/algorithm-main-stages_cat-to-numberic). ### [ctr\_target\_border\_count](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/ctr#ctr_target_border_count) The maximum number of borders to use in target quantization for categorical features that need it. Allowed values are integers from 1 to 255 inclusively. ### [counter\_calc\_method](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/ctr#counter_calc_method) The method for calculating the Counter CTR type. ### [max\_ctr\_complexity](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/ctr#max_ctr_complexity) The maximum number of features that can be combined. ### [ctr\_leaf\_count\_limit](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/ctr#ctr_leaf_count_limit) The maximum number of leaves with categorical features. If the quantity exceeds the specified value a part of leaves is discarded. ### [store\_all\_simple\_ctr](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/ctr#store_all_simple_ctr) Ignore categorical features, which are not used in feature combinations, when choosing candidates for exclusion. ### [final\_ctr\_computation\_mode](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/ctr#final_ctr_computation_mode) Final CTR computation mode. ## Input file settings ### [\-f, --learn-set](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/input#-f) The path to the input file that contains the dataset description. ### [\-t, --test-set](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/input#-t) A comma-separated list of input files that contain the validation dataset description (the format must be the same as used in the training dataset). ### [\--cd, --column-description](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/input#--cd) The path to the input file that contains the [columns description](https://catboost.ai/docs/en/references/training-parameters/en/concepts/input-data_column-descfile). ### [\--learn-pairs](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/input#--learn-pairs) The path to the input file that contains the [pairs description](https://catboost.ai/docs/en/references/training-parameters/en/concepts/input-data_pairs-description) for the training dataset. ### [\--test-pairs](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/input#--test-pairs) The path to the input file that contains the [pairs description](https://catboost.ai/docs/en/references/training-parameters/en/concepts/input-data_pairs-description) for the validation dataset. ### [\--learn-group-weights](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/input#--learn-group-weights) The path to the input file that contains the weights of groups. Refer to the [Group weights](https://catboost.ai/docs/en/references/training-parameters/en/concepts/input-data_group-weights) section for format details. ### [\--test-group-weights](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/input#--test-group-weights) The path to the input file that contains the weights of groups for the validation dataset. ### [\--learn-baseline](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/input#--learn-baseline) The path to the input file that contains baseline values for the training dataset. ### [\--test-baseline](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/input#--test-baseline) The path to the input file that contains baseline values for the validation dataset. ### [\--delimiter](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/input#--delimiter) The delimiter character used to separate the data in the dataset description input file. ### [\--has-header](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/input#--has-header) Read the column names from the first line of the dataset description file if this parameter is set. ### [\--params-files](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/input#--params-files) The path to the input JSON file that contains the training parameters, for example: ### [\--nan-mode](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/input#--nan-mode) The method for [processing missing values](https://catboost.ai/docs/en/references/training-parameters/en/concepts/algorithm-missing-values-processing) in the input dataset. ## Multiclassification settings ### [classes\_count](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/multiclassification#classes_count) Command-line: `--classes-count` The upper limit for the numeric class label. Defines the number of classes for multiclassification. ### [\--class-names](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/multiclassification#--class-names) This parameter is only for Command-line. Classes names. Allows to redefine the default values when using the MultiClass and Logloss metrics. ## Output settings ### [logging\_level](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/output#logging_level) Command line: `--logging-level` The logging level to output to stdout. ### [metric\_period](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/output#metric_period) Command line: `--metric-period` The frequency of iterations to calculate the values of [objectives and metrics](https://catboost.ai/docs/en/references/training-parameters/en/concepts/loss-functions). The usage of this parameter speeds up the training. ### [verbose](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/output#verbose) Command line: `--verbose` Alias:`verbose_eval` The purpose of this parameter depends on the type of the given value: ### [train\_dir](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/output#train_dir) Command line: `--train-dir` The directory for storing the files generated during training. ### [model\_size\_reg](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/output#model_size_reg) Command line: `--model-size-reg` The model size regularization coefficient. The larger the value, the smaller the model size. Refer to the [Model size regularization coefficient](https://catboost.ai/docs/en/references/training-parameters/en/references/model-size-reg) section for details. This regularization is needed only for models with categorical features (other models are small). ### [allow\_writing\_files](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/output#allow_writing_files) Allow to write analytical and snapshot files during training. ### [save\_snapshot](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/output#save_snapshot) Enable snapshotting for [restoring the training progress after an interruption](https://catboost.ai/docs/en/references/training-parameters/en/features/snapshots). ### [snapshot\_file](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/output#snapshot_file) The name of the file to save the training progress information in. This file is used for [recovering training after an interruption](https://catboost.ai/docs/en/references/training-parameters/en/features/snapshots). ### [snapshot\_interval](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/output#snapshot_interval) The interval between saving snapshots in seconds. ### [roc\_file](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/output#roc_file) The name of the [output file](https://catboost.ai/docs/en/references/training-parameters/en/concepts/output-data_roc-curve-points) to save the ROC curve points to. ## Overfitting detection settings ### [early\_stopping\_rounds](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/overfitting-detection#early_stopping_rounds) Sets the overfitting detector type to Iter and stops the training after the specified number of iterations since the iteration with the optimal metric value. ### [od\_type](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/overfitting-detection#od_type) Command-line: `--od-type` The type of the overfitting detector to use. ### [od\_pval](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/overfitting-detection#od_pval) Command-line: `--od-pval` The threshold for the IncToDec [overfitting detector](https://catboost.ai/docs/en/references/training-parameters/en/concepts/overfitting-detector) type. ### [od\_wait](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/overfitting-detection#od_wait) Command-line: `--od-wait` The number of iterations to continue the training after the iteration with the optimal metric value. ## Performance settings ### [thread\_count](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/performance#thread_count) Command-line: `-T`, `--thread-count` The number of threads to use during the training. ### [used\_ram\_limit](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/performance#used_ram_limit) Command-line: `--used-ram-limit` Attempt to limit the amount of used CPU RAM. ### [gpu\_ram\_part](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/performance#gpu_ram_part) Command-line: `--gpu-ram-part` How much of the GPU RAM to use for training. ### [pinned\_memory\_size](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/performance#pinned_memory_size) Command-line: `--pinned-memory-size` How much pinned (page-locked) CPU RAM to use per GPU. ### [gpu\_cat\_features\_storage](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/performance#gpu_cat_features_storage) Command-line: `--gpu-cat-features-storage` The method for storing the categorical features' values. ### [data\_partition](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/performance#data_partition) Command-line: `--data-partition` The method for splitting the input dataset between multiple workers. ## Processing unit settings ### [task\_type](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/processing-unit#task_type) Command line: `--task-type` The processing unit type to use for training. ### [devices](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/#devices) Command line: `--devices` IDs of the GPU devices to use for training (indices are zero-based). ## Quantization settings ### [target\_border](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/quantization#target_border) Command-line: `--target-border` If set, defines the border for converting target values to 0 and 1. ### [border\_count](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/quantization#border_count) Command-line: `-x`, `--border-count` Alias: `max_bin` The number of splits for numerical features. Allowed values are integers from 1 to 65535 inclusively. ### [feature\_border\_type](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/quantization#feature_border_type) Command-line: `--feature-border-type` The [quantization mode](https://catboost.ai/docs/en/references/training-parameters/en/concepts/quantization) for numerical features. ### [per\_float\_feature\_quantization](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/quantization#per_float_feature_quantization) Command-line: `--per-float-feature-quantization` The quantization description for the specified feature or list of features. ## Text processing parameters These parameters are only for the Python package and Command-line version. ### [tokenizers](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/text-processing#tokenizers) Command-line: `--tokenizers` Tokenizers used to preprocess Text type feature columns before creating the dictionary. ### [dictionaries](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/text-processing#dictionaries) Command-line: `--dictionaries` Dictionaries used to preprocess Text type feature columns. Format: ### [feature\_calcers](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/text-processing#feature_calcers) Command-line: `--feature-calcers` Feature calcers used to calculate new features based on preprocessed Text type feature columns. Format: ### [text\_processing](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/text-processing#text_processing) Command-line: `--text-processing` A JSON specification of tokenizers, dictionaries and feature calcers, which determine how text features are converted into a list of float features. [Example](https://catboost.ai/docs/en/references/training-parameters/en/references/text-processing__specification-example) Refer to the description of the following parameters for details on supported values: ## Visualization settings These parameters are only for the Python package. ### [name](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/visualization#name) The experiment name to display in [visualization tools](https://catboost.ai/docs/en/references/training-parameters/en/features/visualization). ### Was the article helpful? Yes No Previous [Object importances](https://catboost.ai/docs/en/references/training-parameters/en/features/object-importances-calcution) Next [Common parameters](https://catboost.ai/docs/en/references/training-parameters/en/references/training-parameters/common) ![](https://mc.yandex.ru/watch/60763294)
Readable Markdown	These parameters are for the Python package, R package and Command-line version. For the Python package several parameters have aliases. For example, the `--iterations` parameter has the following synonyms: `num_boost_round`, `n_estimators`, `num_trees`. Simultaneous usage of different names of one parameter raises an error. Training or inference on CUDA-enabled GPUs requires NVIDIA Driver of version 450.80.02 or higher. ## Common parameters ### [loss\_function](https://catboost.ai/docs/en/references/training-parameters/common#loss_function) Command-line: `--loss-function` Alias: `objective` The [metric](https://catboost.ai/docs/en/concepts/loss-functions) to use in training. The specified value also determines the machine learning problem to solve. Some metrics support optional parameters (see the [Objectives and metrics](https://catboost.ai/docs/en/concepts/loss-functions) section for details on each metric). ### [custom\_metric](https://catboost.ai/docs/en/references/training-parameters/common#custom_metric) Command-line: `--custom-metric` [Metric](https://catboost.ai/docs/en/concepts/loss-functions) values to output during training. These functions are not optimized and are displayed for informational purposes only. Some metrics support optional parameters (see the [Objectives and metrics](https://catboost.ai/docs/en/concepts/loss-functions) section for details on each metric). ### [eval\_metric](https://catboost.ai/docs/en/references/training-parameters/common#eval_metric) Command-line: `--eval-metric` The metric used for overfitting detection (if enabled) and best model selection (if enabled). Some metrics support optional parameters (see the [Objectives and metrics](https://catboost.ai/docs/en/concepts/loss-functions) section for details on each metric). ### [iterations](https://catboost.ai/docs/en/references/training-parameters/common#iterations) Command-line: `-i`, `--iterations` Aliases: `num_boost_round`, `n_estimators`, `num_trees` The maximum number of trees that can be built when solving machine learning problems. ### [learning\_rate](https://catboost.ai/docs/en/references/training-parameters/common#learning_rate) Command-line: `-w`, `--learning-rate` Alias: `eta` The learning rate. Used for reducing the gradient step. ### [random\_seed](https://catboost.ai/docs/en/references/training-parameters/common#random_seed) Command-line: `-r`, `--random-seed` Alias:`random_state` The random seed used for training. ### [l2\_leaf\_reg](https://catboost.ai/docs/en/references/training-parameters/common#l2_leaf_reg) Command-line: `--l2-leaf-reg`, `l2-leaf-regularizer` Alias: `reg_lambda` Coefficient at the L2 regularization term of the cost function. ### [bootstrap\_type](https://catboost.ai/docs/en/references/training-parameters/common#bootstrap_type) Command-line: `--bootstrap-type` [Bootstrap type](https://catboost.ai/docs/en/concepts/algorithm-main-stages_bootstrap-options). Defines the method for sampling the weights of objects. ### [bagging\_temperature](https://catboost.ai/docs/en/references/training-parameters/common#bagging_temperature) Command-line: `--bagging-temperature` Defines the settings of the Bayesian bootstrap. It is used by default in classification and regression modes. ### [subsample](https://catboost.ai/docs/en/references/training-parameters/common#subsample) Command-line: `--subsample` Sample rate for bagging. ### [sampling\_frequency](https://catboost.ai/docs/en/references/training-parameters/common#sampling_frequency) Command-line: `--sampling-frequency` Frequency to sample weights and objects when building trees. ### [sampling\_unit](https://catboost.ai/docs/en/references/training-parameters/common#sampling_unit) Command-line: `--sampling-unit` The sampling scheme. ### [mvs\_reg](https://catboost.ai/docs/en/references/training-parameters/common#mvs_reg) Command-line: `--mvs-reg` Affects the weight of the denominator and can be used for balancing between the importance and Bernoulli sampling (setting it to 0 implies importance sampling and to ∞ \\infty - Bernoulli). ### [random\_strength](https://catboost.ai/docs/en/references/training-parameters/common#random_strength) Command-line: `--random-strength` The amount of randomness to use for scoring splits when the tree structure is selected. Use this parameter to avoid overfitting the model. ### [use\_best\_model](https://catboost.ai/docs/en/references/training-parameters/common#use_best_model) Command-line: `--use-best-model` Use the validation dataset to identify the iteration with the optimal value of the evaluation metric. Save only trees up to this iteration (inclusive) in the resulting model. ### [best\_model\_min\_trees](https://catboost.ai/docs/en/references/training-parameters/common#best_model_min_trees) Command-line: `--best-model-min-trees` The minimal number of trees that the best model should have. If set, the output model contains at least the given number of trees even if the optimal value of the evaluation metric on the validation dataset is achieved with smaller number of trees. ### [depth](https://catboost.ai/docs/en/references/training-parameters/common#depth) Command-line: `-n`, `--depth` Alias: `max_depth` Depth of the trees. ### [grow\_policy](https://catboost.ai/docs/en/references/training-parameters/common#grow_policy) Command-line: `--grow-policy` The tree growing policy. Defines how to perform greedy tree construction. ### [min\_data\_in\_leaf](https://catboost.ai/docs/en/references/training-parameters/common#min_data_in_leaf) Command-line: `--min-data-in-leaf` Alias: `min_child_samples` The minimum number of training samples in a leaf. CatBoost does not search for new splits in leaves with samples count less than the specified value. ### [max\_leaves](https://catboost.ai/docs/en/references/training-parameters/common#max_leavescommon.md#) Command-line: `--max-leaves` Alias:`num_leaves` The maximum number of leafs in the resulting tree. Can be used only with the Lossguide growing policy. ### [ignored\_features](https://catboost.ai/docs/en/references/training-parameters/common#ignored_features) Command-line: `-I`, `--ignore-features` Feature indices or names to exclude from the training. It is assumed that all passed values are feature names if at least one of the passed values can not be converted to a number or a range of numbers. Otherwise, it is assumed that all passed values are feature indices. Specifics: - Non-negative indices that do not match any features are successfully ignored. For example, if five features are defined for the objects in the dataset and this parameter is set to "42", the corresponding non-existing feature is successfully ignored. - The identifier corresponds to the feature's index. Feature indices used in train and feature importance are numbered from 0 to `featureCount – 1`. If a file is used as [input data](https://catboost.ai/docs/en/concepts/input-data) then any non-feature column types are ignored when calculating these indices. For example, each row in the input file contains data in the following order: `cat feature<\t>label value<\t>num feature`. So for the row `rock<\t>0<\t>42`, the identifier for the "rock" feature is 0, and for the "42" feature it's 1. - The addition of a non-existing feature name raises an error. ### [one\_hot\_max\_size](https://catboost.ai/docs/en/references/training-parameters/common#one_hot_max_size) Command-line: `--one-hot-max-size` Use one-hot encoding for all categorical features with a number of different values less than or equal to the given parameter value. Ctrs are not calculated for such features. ### [has\_time](https://catboost.ai/docs/en/references/training-parameters/common#has_time) Command-line: `--has-time` Use the order of objects in the input data (do not perform random permutations during the [Transforming categorical features to numerical features](https://catboost.ai/docs/en/concepts/algorithm-main-stages_cat-to-numberic) and [Choosing the tree structure](https://catboost.ai/docs/en/concepts/algorithm-main-stages_choose-tree-structure) stages). ### [rsm](https://catboost.ai/docs/en/references/training-parameters/common#rsm) Command-line: `--rsm` Alias:`colsample_bylevel` Random subspace method. The percentage of features to use at each split selection, when features are selected over again at random. ### [nan\_mode](https://catboost.ai/docs/en/references/training-parameters/common#nan_mode) Command-line: `--nan-mode` The method for [processing missing values](https://catboost.ai/docs/en/concepts/algorithm-missing-values-processing) in the input dataset. ### [input\_borders](https://catboost.ai/docs/en/references/training-parameters/common#input_borders) Command-line: `--input-borders-file` Load [Custom quantization borders and missing value modes](https://catboost.ai/docs/en/concepts/input-data_custom-borders) from a file (do not generate them). ### [output\_borders](https://catboost.ai/docs/en/references/training-parameters/common#output_borders) Command-line: `--output-borders-file` Save quantization borders for the current dataset to a file. ### [fold\_permutation\_block](https://catboost.ai/docs/en/references/training-parameters/common#fold_permutation_block) Command-line: `--fold-permutation-block` Objects in the dataset are grouped in blocks before the random permutations. This parameter defines the size of the blocks. ### [leaf\_estimation\_method](https://catboost.ai/docs/en/references/training-parameters/common#leaf_estimation_method) Command-line: `--leaf-estimation-method` The method used to calculate the values in leaves. ### [leaf\_estimation\_iterations](https://catboost.ai/docs/en/references/training-parameters/common#leaf_estimation_iterations) Command-line: `--leaf-estimation-iterations` This parameter regulates how many steps are done in every tree when calculating leaf values. ### [leaf\_estimation\_backtracking](https://catboost.ai/docs/en/references/training-parameters/common#leaf_estimation_backtracking) Command-line: `--leaf-estimation-backtracking` When the value of the `leaf_estimation_iterations` parameter is greater than 1, CatBoost makes several gradient or newton steps when calculating the resulting leaf values of a tree. ### [fold\_len\_multiplier](https://catboost.ai/docs/en/references/training-parameters/common#fold_len_multiplier) Command-line: `--fold-len-multiplier` Coefficient for changing the length of folds. ### [approx\_on\_full\_history](https://catboost.ai/docs/en/references/training-parameters/common#approx_on_full_history) Command-line:`--approx-on-full-history` The principles for calculating the approximated values. ### [class\_weights](https://catboost.ai/docs/en/references/training-parameters/common#class_weights) Command-line: `--class-weights` Class weights. The values are used as multipliers for the object weights. This parameter can be used for solving binary classification and multiclassification problems. ### [class\_names](https://catboost.ai/docs/en/references/training-parameters/common#class_names) Classes names. Allows to redefine the default values when using the MultiClass and Logloss metrics. ### [auto\_class\_weights](https://catboost.ai/docs/en/references/training-parameters/common#auto_class_weights) Command-line: `--auto-class-weights` Automatically calculate class weights based either on the total weight or the total number of objects in each class. The values are used as multipliers for the object weights. Supported values: - None — All class weights are set to 1 - Balanced: C W k \= m a x c \= 1 K ( ∑ t i \= c w i ) ∑ t i \= k w i CW\_k=\\displaystyle\\frac{max\_{c=1}^K(\\sum\_{t\_{i}=c}{w\_i})}{\\sum\_{t\_{i}=k}{w\_{i}}} - SqrtBalanced: C W k \= m a x c \= 1 K ( ∑ t i \= c w i ) ∑ t i \= k w i CW\_k=\\sqrt{\\displaystyle\\frac{max\_{c=1}^K(\\sum\_{t\_i=c}{w\_i})}{\\sum\_{t\_i=k}{w\_i}}} ### [scale\_pos\_weight](https://catboost.ai/docs/en/references/training-parameters/common#scale_pos_weight) The weight for class 1 in binary classification. The value is used as a multiplier for the weights of objects from class 1. ### [boosting\_type](https://catboost.ai/docs/en/references/training-parameters/common#boosting_type) Command-line: `--boosting-type` Boosting scheme. ### [boost\_from\_average](https://catboost.ai/docs/en/references/training-parameters/common#boost_from_average) Command-line: `--boost-from-average` Initialize approximate values by best constant value for the specified loss function. ### [langevin](https://catboost.ai/docs/en/references/training-parameters/common#langevin) Command-line: `--langevin` Enables the Stochastic Gradient Langevin Boosting mode. ### [diffusion\_temperature](https://catboost.ai/docs/en/references/training-parameters/common#diffusion_temperature) Command-line: `--diffusion-temperature` The diffusion temperature of the Stochastic Gradient Langevin Boosting mode. ### [posterior\_sampling](https://catboost.ai/docs/en/references/training-parameters/common#posterior_sampling) Command-line: `--posterior-sampling` If this parameter is set several options are specified as follows and model parameters are checked to obtain uncertainty predictions with good theoretical properties. ### [allow\_const\_label](https://catboost.ai/docs/en/references/training-parameters/common#allow_const_label) Command-line: `--allow-const-label` Use it to train models with datasets that have equal label values for all objects. ### [score\_function](https://catboost.ai/docs/en/references/training-parameters/common#score_function) Command-line: `--score-function` The [score type](https://catboost.ai/docs/en/concepts/algorithm-score-functions) used to select the next split during the tree construction. ### [monotone\_constraints](https://catboost.ai/docs/en/references/training-parameters/common#monotone_constraints) Command-line: `--monotone-constraints` Impose monotonic constraints on numerical features. Possible values: - "1" — Increasing constraint on the feature. The algorithm forces the model to be a non-decreasing function of this features. - "\-1" — Decreasing constraint on the feature. The algorithm forces the model to be a non-increasing function of this features. - "0" — constraints are disabled. ### [feature\_weights](https://catboost.ai/docs/en/references/training-parameters/common#feature_weights) Command-line: `--feature-weights` Per-feature multiplication weights used when choosing the best split. The score of each candidate is multiplied by the weights of features from the current split. Non-negative float values are supported for each weight. Supported formats for setting the value of this parameter: ### [first\_feature\_use\_penalties](https://catboost.ai/docs/en/references/training-parameters/common#first_feature_use_penalties) Command-line: `--first-feature-use-penalties` Per-feature penalties for the first occurrence of the feature in the model. The given value is subtracted from the score if the current candidate is the first one to include the feature in the model. Refer to the [Per-object and per-feature penalties](https://catboost.ai/docs/en/concepts/algorithm-score-functions) section for details on applying different score penalties. Non-negative float values are supported for each penalty. ### [fixed\_binary\_splits](https://catboost.ai/docs/en/references/training-parameters/common#fixed_binary_splits) Command-line: `--fixed-binary-splits` A list of indices of binary features to put at the top of each tree. ### [penalties\_coefficient](https://catboost.ai/docs/en/references/training-parameters/common#penalties_coefficient) Command-line: `--penalties-coefficient` A single-value common coefficient to multiply all penalties. ### [per\_object\_feature\_penalties](https://catboost.ai/docs/en/references/training-parameters/common#per_object_feature_penalties) Command-line: `--per-object-feature-penalties` Per-object penalties for the first use of the feature for the object. The given value is multiplied by the number of objects that are divided by the current split and use the feature for the first time. Refer to the [Per-object and per-feature penalties](https://catboost.ai/docs/en/concepts/algorithm-score-functions) section for details on applying different score penalties. Non-negative float values are supported for each penalty. ### [model\_shrink\_rate](https://catboost.ai/docs/en/references/training-parameters/common#model_shrink_rate) Command-line: `--model-shrink-rate` The constant used to calculate the coefficient for multiplying the model on each iteration. ### [model\_shrink\_mode](https://catboost.ai/docs/en/references/training-parameters/common#model_shrink_mode) Command-line: `model_shrink_mode` Determines how the actual model shrinkage coefficient is calculated at each iteration. ## CTR settings ### [simple\_ctr](https://catboost.ai/docs/en/references/training-parameters/ctr#simple_ctr) Quantization settings for simple [categorical features](https://catboost.ai/docs/en/concepts/algorithm-main-stages_cat-to-numberic). Use this parameter to specify the principles for defining the class of the object for regression tasks. By default, it is considered that an object belongs to the positive class if its' label value is greater than the median of all label values of the dataset. ### [combinations\_ctr](https://catboost.ai/docs/en/references/training-parameters/ctr#combinations_ctr) Quantization settings for combinations of [categorical features](https://catboost.ai/docs/en/concepts/algorithm-main-stages_cat-to-numberic). ### [per\_feature\_ctr](https://catboost.ai/docs/en/references/training-parameters/ctr#per_feature_ctr) Per-feature quantization settings for [categorical features](https://catboost.ai/docs/en/concepts/algorithm-main-stages_cat-to-numberic). ### [ctr\_target\_border\_count](https://catboost.ai/docs/en/references/training-parameters/ctr#ctr_target_border_count) The maximum number of borders to use in target quantization for categorical features that need it. Allowed values are integers from 1 to 255 inclusively. ### [counter\_calc\_method](https://catboost.ai/docs/en/references/training-parameters/ctr#counter_calc_method) The method for calculating the Counter CTR type. ### [max\_ctr\_complexity](https://catboost.ai/docs/en/references/training-parameters/ctr#max_ctr_complexity) The maximum number of features that can be combined. ### [ctr\_leaf\_count\_limit](https://catboost.ai/docs/en/references/training-parameters/ctr#ctr_leaf_count_limit) The maximum number of leaves with categorical features. If the quantity exceeds the specified value a part of leaves is discarded. ### [store\_all\_simple\_ctr](https://catboost.ai/docs/en/references/training-parameters/ctr#store_all_simple_ctr) Ignore categorical features, which are not used in feature combinations, when choosing candidates for exclusion. ### [final\_ctr\_computation\_mode](https://catboost.ai/docs/en/references/training-parameters/ctr#final_ctr_computation_mode) Final CTR computation mode. ## Input file settings ### [\-f, --learn-set](https://catboost.ai/docs/en/references/training-parameters/input#-f) The path to the input file that contains the dataset description. ### [\-t, --test-set](https://catboost.ai/docs/en/references/training-parameters/input#-t) A comma-separated list of input files that contain the validation dataset description (the format must be the same as used in the training dataset). ### [\--cd, --column-description](https://catboost.ai/docs/en/references/training-parameters/input#--cd) The path to the input file that contains the [columns description](https://catboost.ai/docs/en/concepts/input-data_column-descfile). ### [\--learn-pairs](https://catboost.ai/docs/en/references/training-parameters/input#--learn-pairs) The path to the input file that contains the [pairs description](https://catboost.ai/docs/en/concepts/input-data_pairs-description) for the training dataset. ### [\--test-pairs](https://catboost.ai/docs/en/references/training-parameters/input#--test-pairs) The path to the input file that contains the [pairs description](https://catboost.ai/docs/en/concepts/input-data_pairs-description) for the validation dataset. ### [\--learn-group-weights](https://catboost.ai/docs/en/references/training-parameters/input#--learn-group-weights) The path to the input file that contains the weights of groups. Refer to the [Group weights](https://catboost.ai/docs/en/concepts/input-data_group-weights) section for format details. ### [\--test-group-weights](https://catboost.ai/docs/en/references/training-parameters/input#--test-group-weights) The path to the input file that contains the weights of groups for the validation dataset. ### [\--learn-baseline](https://catboost.ai/docs/en/references/training-parameters/input#--learn-baseline) The path to the input file that contains baseline values for the training dataset. ### [\--test-baseline](https://catboost.ai/docs/en/references/training-parameters/input#--test-baseline) The path to the input file that contains baseline values for the validation dataset. ### [\--delimiter](https://catboost.ai/docs/en/references/training-parameters/input#--delimiter) The delimiter character used to separate the data in the dataset description input file. Read the column names from the first line of the dataset description file if this parameter is set. ### [\--params-files](https://catboost.ai/docs/en/references/training-parameters/input#--params-files) The path to the input JSON file that contains the training parameters, for example: ### [\--nan-mode](https://catboost.ai/docs/en/references/training-parameters/input#--nan-mode) The method for [processing missing values](https://catboost.ai/docs/en/concepts/algorithm-missing-values-processing) in the input dataset. ## Multiclassification settings ### [classes\_count](https://catboost.ai/docs/en/references/training-parameters/multiclassification#classes_count) Command-line: `--classes-count` The upper limit for the numeric class label. Defines the number of classes for multiclassification. ### [\--class-names](https://catboost.ai/docs/en/references/training-parameters/multiclassification#--class-names) This parameter is only for Command-line. Classes names. Allows to redefine the default values when using the MultiClass and Logloss metrics. ## Output settings ### [logging\_level](https://catboost.ai/docs/en/references/training-parameters/output#logging_level) Command line: `--logging-level` The logging level to output to stdout. ### [metric\_period](https://catboost.ai/docs/en/references/training-parameters/output#metric_period) Command line: `--metric-period` The frequency of iterations to calculate the values of [objectives and metrics](https://catboost.ai/docs/en/concepts/loss-functions). The usage of this parameter speeds up the training. ### [verbose](https://catboost.ai/docs/en/references/training-parameters/output#verbose) Command line: `--verbose` Alias:`verbose_eval` The purpose of this parameter depends on the type of the given value: ### [train\_dir](https://catboost.ai/docs/en/references/training-parameters/output#train_dir) Command line: `--train-dir` The directory for storing the files generated during training. ### [model\_size\_reg](https://catboost.ai/docs/en/references/training-parameters/output#model_size_reg) Command line: `--model-size-reg` The model size regularization coefficient. The larger the value, the smaller the model size. Refer to the [Model size regularization coefficient](https://catboost.ai/docs/en/references/model-size-reg) section for details. This regularization is needed only for models with categorical features (other models are small). ### [allow\_writing\_files](https://catboost.ai/docs/en/references/training-parameters/output#allow_writing_files) Allow to write analytical and snapshot files during training. ### [save\_snapshot](https://catboost.ai/docs/en/references/training-parameters/output#save_snapshot) Enable snapshotting for [restoring the training progress after an interruption](https://catboost.ai/docs/en/features/snapshots). ### [snapshot\_file](https://catboost.ai/docs/en/references/training-parameters/output#snapshot_file) The name of the file to save the training progress information in. This file is used for [recovering training after an interruption](https://catboost.ai/docs/en/features/snapshots). ### [snapshot\_interval](https://catboost.ai/docs/en/references/training-parameters/output#snapshot_interval) The interval between saving snapshots in seconds. ### [roc\_file](https://catboost.ai/docs/en/references/training-parameters/output#roc_file) The name of the [output file](https://catboost.ai/docs/en/concepts/output-data_roc-curve-points) to save the ROC curve points to. ## Overfitting detection settings ### [early\_stopping\_rounds](https://catboost.ai/docs/en/references/training-parameters/overfitting-detection#early_stopping_rounds) Sets the overfitting detector type to Iter and stops the training after the specified number of iterations since the iteration with the optimal metric value. ### [od\_type](https://catboost.ai/docs/en/references/training-parameters/overfitting-detection#od_type) Command-line: `--od-type` The type of the overfitting detector to use. ### [od\_pval](https://catboost.ai/docs/en/references/training-parameters/overfitting-detection#od_pval) Command-line: `--od-pval` The threshold for the IncToDec [overfitting detector](https://catboost.ai/docs/en/concepts/overfitting-detector) type. ### [od\_wait](https://catboost.ai/docs/en/references/training-parameters/overfitting-detection#od_wait) Command-line: `--od-wait` The number of iterations to continue the training after the iteration with the optimal metric value. ## Performance settings ### [thread\_count](https://catboost.ai/docs/en/references/training-parameters/performance#thread_count) Command-line: `-T`, `--thread-count` The number of threads to use during the training. ### [used\_ram\_limit](https://catboost.ai/docs/en/references/training-parameters/performance#used_ram_limit) Command-line: `--used-ram-limit` Attempt to limit the amount of used CPU RAM. ### [gpu\_ram\_part](https://catboost.ai/docs/en/references/training-parameters/performance#gpu_ram_part) Command-line: `--gpu-ram-part` How much of the GPU RAM to use for training. ### [pinned\_memory\_size](https://catboost.ai/docs/en/references/training-parameters/performance#pinned_memory_size) Command-line: `--pinned-memory-size` How much pinned (page-locked) CPU RAM to use per GPU. ### [gpu\_cat\_features\_storage](https://catboost.ai/docs/en/references/training-parameters/performance#gpu_cat_features_storage) Command-line: `--gpu-cat-features-storage` The method for storing the categorical features' values. ### [data\_partition](https://catboost.ai/docs/en/references/training-parameters/performance#data_partition) Command-line: `--data-partition` The method for splitting the input dataset between multiple workers. ## Processing unit settings ### [task\_type](https://catboost.ai/docs/en/references/training-parameters/processing-unit#task_type) Command line: `--task-type` The processing unit type to use for training. ### [devices](https://catboost.ai/docs/en/references/training-parameters/#devices) Command line: `--devices` IDs of the GPU devices to use for training (indices are zero-based). ## Quantization settings ### [target\_border](https://catboost.ai/docs/en/references/training-parameters/quantization#target_border) Command-line: `--target-border` If set, defines the border for converting target values to 0 and 1. ### [border\_count](https://catboost.ai/docs/en/references/training-parameters/quantization#border_count) Command-line: `-x`, `--border-count` Alias: `max_bin` The number of splits for numerical features. Allowed values are integers from 1 to 65535 inclusively. ### [feature\_border\_type](https://catboost.ai/docs/en/references/training-parameters/quantization#feature_border_type) Command-line: `--feature-border-type` The [quantization mode](https://catboost.ai/docs/en/concepts/quantization) for numerical features. ### [per\_float\_feature\_quantization](https://catboost.ai/docs/en/references/training-parameters/quantization#per_float_feature_quantization) Command-line: `--per-float-feature-quantization` The quantization description for the specified feature or list of features. ## Text processing parameters These parameters are only for the Python package and Command-line version. ### [tokenizers](https://catboost.ai/docs/en/references/training-parameters/text-processing#tokenizers) Command-line: `--tokenizers` Tokenizers used to preprocess Text type feature columns before creating the dictionary. ### [dictionaries](https://catboost.ai/docs/en/references/training-parameters/text-processing#dictionaries) Command-line: `--dictionaries` Dictionaries used to preprocess Text type feature columns. Format: ### [feature\_calcers](https://catboost.ai/docs/en/references/training-parameters/text-processing#feature_calcers) Command-line: `--feature-calcers` Feature calcers used to calculate new features based on preprocessed Text type feature columns. Format: ### [text\_processing](https://catboost.ai/docs/en/references/training-parameters/text-processing#text_processing) Command-line: `--text-processing` A JSON specification of tokenizers, dictionaries and feature calcers, which determine how text features are converted into a list of float features. [Example](https://catboost.ai/docs/en/references/text-processing__specification-example) Refer to the description of the following parameters for details on supported values: ## Visualization settings These parameters are only for the Python package. ### [name](https://catboost.ai/docs/en/references/training-parameters/visualization#name) The experiment name to display in [visualization tools](https://catboost.ai/docs/en/features/visualization).
Shard	169 (laksa)
Root Hash	17435841955170310369
Unparsed URL	ai,catboost!/docs/en/references/training-parameters/ s443