1.2.0

1.2.0

Bug Fixes

• Model Registry: Fix "XGBoost version not compiled with GPU support" error when running CPU inference against open-source
  XGBoost models deployed to SPCS.
• Model Registry: Fix model deployment to SPCS on Windows machines.

Behavior Changes

New Features

• Model Development: Introduced XGBoost external memory training feature. This feature enables training XGBoost models
  on large datasets that don't fit into memory.
• Registry: New Registry class named snowflake.ml.registry.Registry providing similar APIs as the old one but works
  with new MODEL object in Snowflake SQL. Also, we are providingsnowflake.ml.model.Model and
  snowflake.ml.model.ModelVersion to represent a model and a specific version of a model.
• Model Development: Add support for fit_predict method in AgglomerativeClustering, DBSCAN, and OPTICS classes;
• Model Development: Add support for fit_transform method in MDS, SpectralEmbedding and TSNE class.

Additional Notes

• Model Registry: The snowflake.ml.registry.model_registry.ModelRegistry has been deprecated starting from version
  1.2.0. It will stay in the Private Preview phase. For future implementations, kindly utilize
  snowflake.ml.registry.Registry, except when specifically required. The old model registry will be removed once all
  its primary functionalities are fully integrated into the new registry.

1.1.2

1.1.2

Bug Fixes

• Generic: Fix the issue that stack trace is hidden by telemetry unexpectedly.
• Model Development: Execute model signature inference without materializing full dataframe in memory.
• Model Registry: Fix occasional 'snowflake-ml-python library does not exist' error when deploying to SPCS.

Behavior Changes

• Model Registry: When calling predict with Snowpark DataFrame, both inferred or normalized column names are accepted.
• Model Registry: When logging a Snowpark ML Modeling Model, sample input data or manually provided signature will be
  ignored since they are not necessary.

New Features

• Model Development: SQL implementation of binary precision_score metric.

1.1.1

1.1.1

Bug Fixes

• Model Registry: The predict target method on registered models is now compatible with unsupervised estimators.
• Model Development: Fix confusion_matrix incorrect results when the row number cannot be divided by the batch size.

Behavior Changes

New Features

• Introduced passthrough_col param in Modeling API. This new param is helpful in scenarios
  requiring automatic input_cols inference, but need to avoid using specific
  columns, like index columns, during training or inference.

1.1.0

1.1.0

Bug Fixes

• Model Registry: Fix panda dataframe input not handling first row properly.
• Model Development: OrdinalEncoder and LabelEncoder output_columns do not need to be valid snowflake identifiers. They
  would previously be excluded if the normalized name did not match the name specified in output_columns.

Behavior Changes

New Features

• Model Registry: Add support for invoking public endpoint on SPCS service, by providing a "enable_ingress" SPCS
  deployment option.
• Model Development: Add support for distributed HPO - GridSearchCV and RandomizedSearchCV execution will be
  distributed on multi-node warehouses.

1.0.12

1.0.12

Bug Fixes

• Model Registry: Fix regression issue that container logging is not shown during model deployment to SPCS.
• Model Development: Enhance the column capacity of OrdinalEncoder.
• Model Registry: Fix unbound `batch_size`` error when deploying a model other than Hugging Face Pipeline
  and LLM with GPU on SPCS.

Behavior Changes

• Model Registry: Raise early error when deploying to SPCS with db/schema that starts with underscore.
• Model Registry: conda-forge channel is now automatically added to channel lists when deploying to SPCS.
• Model Registry: relax_version will not strip all version specifier, instead it will relax ==x.y.z specifier to
  >=x.y,<(x+1).
• Model Registry: Python with different patchlevel but the same major and minor will not result a warning when loading
  the model via Model Registry and would be considered to use when deploying to SPCS.
• Model Registry: When logging a snowflake.ml.model.models.huggingface_pipeline.HuggingFacePipelineModel object,
  versions of local installed libraries won't be picked as dependencies of models, instead it will pick up some pre-
  defined dependencies to improve user experience.

New Features

• Model Registry: Enable best-effort SPCS job/service log streaming when logging level is set to INFO.

1.0.11

1.0.11

New Features

• Model Registry: Add log_artifact() public method.
• Model Development: Add support for kneighbors.

Behavior Changes

• Model Registry: Change log_model() argument from TrainingDataset to List of Artifact.
• Model Registry: Change get_training_dataset() to get_artifact().

Bug Fixes

• Model Development: Fix support for XGBoost and LightGBM models using SKLearn Grid Search and Randomized Search model selectors.
• Model Development: DecimalType is now supported as a DataType.
• Model Development: Fix metrics compatibility with Snowpark Dataframes that use Snowflake identifiers

1.0.10

1.0.10

Behavior Changes

• Model Development: precision_score, recall_score, f1_score, fbeta_score, precision_recall_fscore_support,
  mean_absolute_error, mean_squared_error, and mean_absolute_percentage_error metric calculations are now distributed.
• Model Registry: deploy will now return Deployment for deployment information.

New Features

• Model Registry: When the model signature is auto-inferred, it will be printed to the log for reference.
• Model Registry: For SPCS deployment, Deployment details will contains image_name, service_spec and service_function_sql.

Bug Fixes

• Model Development: Fix an issue that leading to UTF-8 decoding errors when using modeling modules on Windows.
• Model Development: Fix an issue that alias definitions cause SnowparkSQLUnexpectedAliasException in inference.
• Model Registry: Fix an issue that signature inference could be incorrect when using Snowpark DataFrame as sample input.
• Model Registry: Fix too strict data type validation when predicting. Now, for example, if you have a INT8
  type feature in the signature, if providing a INT64 dataframe but all values are within the range, it would not fail.

[1.0.9]

Behavior Changes

• Model Development: log_loss metric calculation is now distributed.

Bug Fixes

• Model Registry: Fix an issue that building images fails with specific docker setup.
• Model Registry: Fix an issue that unable to embed local ML library when the library is imported by zipimport.
• Model Registry: Fix out-of-date doc about platform argument in the deploy function.
• Model Registry: Fix an issue that unable to deploy a GPU-trained PyTorch model to a platform where GPU is not available.

[1.0.8]

1.0.8

Bug Fixes

• Model Development: Ordinal encoder can be used with mixed input column types.
• Model Registry: Fix an issue that incorrect docker executable is used when building images.
• Model Registry: Fix an issue that specifying token argument when using
  snowflake.ml.model.models.huggingface_pipeline.HuggingFacePipelineModel with transformers < 4.32.0 is not effective.
• Model Registry: Fix an issue that incorrect system function call is used when deploying to SPCS.
• Model Registry: Fix an issue when using a transformers.pipeline that does not have a tokenizer.
• Model Registry: Fix incorrectly-inferred image repository name during model deployment to SPCS.
• Model Registry: Fix GPU resource retention issue caused by failed or stuck previous deployments in SPCS.

[1.0.7]

Bug Fixes

• Model Development & Model Registry: Fix an error related to pandas.io.json.json_normalize.