Snowflake ML Python release notes¶

See Snowflake ML: End-to-End Agentic Machine Learning for documentation.

Verifying the snowflake-ml-python package¶

All Snowflake packages are signed, allowing you to verify their origin. To verify the snowflake.ml.python package, follow the steps below:

1. Install cosign. This example uses the Go installation: Installing cosign with Go.
2. Download the file from a repository such as PyPi.
3. Download a .sig file for that release from the GitHub releases page.
4. Verify the signature using cosign. For example:

cosign verify-blob snowflake_ml_python-1.27.0.tar.gz --key snowflake-ml-python-1.27.0.pub --signature resources.linux.snowflake_ml_python-1.27.0.tar.gz.sig

cosign verify-blob snowflake_ml_python-1.27.0.tar.gz --key snowflake-ml-python-1.27.0.pub --signature resources.linux.snowflake_ml_python-1.27.0.tar.gz.sig

Deprecation notices¶

• snowflake.ml.fileset.FileSet has been deprecated and will be removed in a future release. Use snowflake.ml.dataset.Dataset and snowflake.ml.data.DataConnector instead.
• The “CamelCase” function names in snowflake.ml.cortex have been deprecated and will be removed in a future release. Use the “snake_case” names for these functions instead. For example, use classify_text instead of ClassifyText.
• The partitioned_inference_api decorator has been deprecated and will be removed in a future release. Use custom_model.partitioned_api instead.
• The additional_payloads argument of the MLJob.submit_* methods has been deprecated and will be removed in a future release. Use the imports argument instead.
• The snowflake.ml.model.models.huggingface_pipeline.HuggingfacePipelineModel class has been deprecated and will be removed in a future release.

Version 1.48.0 (2026-07-22)¶

New Features¶

• Experiment Tracking: ExperimentTracking now captures best-effort source provenance (entry-point filename and any surrounding git commit, branch, and remote URL) for each new run by default. Pass capture_source_info=False to disable it. Collection is always non-fatal and never blocks run creation.

Bug Fixes¶

Behavior Changes¶

Deprecations¶

Version 1.47.0 (2026-07-15)¶

New Features¶

• Experiment Tracking: Added ExperimentTracking.list_model_versions to retrieve the model versions that were logged under a run. Models logged via log_model inside a run are linked to that run through Snowflake lineage, and this method traverses that lineage to return the corresponding ModelVersion objects.

from snowflake.ml.experiment import ExperimentTracking

exp = ExperimentTracking(session)
exp.set_experiment("MY_EXPERIMENT")
with exp.start_run(run_name="MY_RUN"):
    exp.log_model(model, model_name="MY_MODEL", sample_input_data=X)

model_versions = exp.list_model_versions(run_name="MY_RUN")

Bug Fixes¶

• Experiment Tracking: Fix a bug where .set_experiment does not recreate an experiment deleted in Snowsight

Behavior Changes¶

• Registry: enable_explainability now defaults to False for all model types when logging a model. Previously it defaulted to True for XGBoost, LightGBM, and CatBoost models and was auto-enabled for supported scikit-learn and Snowpark ML modeling models when running in the Warehouse. To generate an explain method, explicitly pass options={"enable_explainability": True} to log_model (this requires sample_input_data for supported model types).

Deprecations¶

Version 1.46.0 (2026-07-07)¶

New Features¶

Bug Fixes¶

Behavior Changes¶

• Registry: For Snowpark ML Pipeline models with explainability enabled, explanations now cover only the columns the final estimator was actually trained on (its input_cols). Untransformed passthrough columns (e.g. raw string columns left in place by an encoder using new output_cols, or the label column) are no longer included in the explain output. This also fixes a failure where such passthrough columns could reach the SHAP explainer and raise an error during log_model(..., enable_explainability=True).

Deprecations¶

Version 1.45.0 (2026-07-01)¶

New Features¶

• Feature Store: FeatureView now supports an initialization_warehouse that is used for the initial build and any subsequent reinitializations of the backing dynamic table (a full scan of the source data), while warehouse continues to drive the lighter incremental refreshes. This mirrors the dynamic table INITIALIZATION_WAREHOUSE knob, lets you pair a larger warehouse for initialization with a smaller one for steady-state refresh, and is also used for the one-time backfill of streaming feature views. It can be set at registration, changed via update_feature_view(initialization_warehouse=...), and is surfaced by list_feature_views(verbose=True).

draft_fv = FeatureView(
    name="F_TRIP",
    entities=[entity],
    feature_df=feature_df,
    refresh_freq="1d",
    warehouse="SMALL_WH",            # incremental refreshes
    initialization_warehouse="LARGE_WH",  # initial build / reinitialization
)
fv = fs.register_feature_view(draft_fv, version="1.0")

• Registry: LLM models deployed with the OpenAI chat signatures now support structured outputs through an optional response_format param matching the OpenAI Chat Completions API ({"type": "json_schema", "json_schema": {"name": "...", "schema": {...}}}), letting callers constrain model output to a JSON Schema.

from pydantic import BaseModel
import pandas as pd

class CityCountry(BaseModel):
    city: str
    country: str

response_format = {
    "type": "json_schema",
    "json_schema": {
        "name": "city_country",
        "schema": CityCountry.model_json_schema(),
    },
}

x_df = pd.DataFrame.from_records(
    [
        {
            "messages": [
                {
                    "role": "user",
                    "content": [
                        {"type": "text", "text": "What is the capital of France?"},
                    ],
                },
            ],
        }
    ]
)

mv.run(
    X=x_df,
    params={"response_format": response_format},
    service_name=...,
)

Bug Fixes¶

Behavior Changes¶

Deprecations¶

Version 1.44.0 (2026-06-23)¶

New Features¶

Bug Fixes¶

Behavior Changes¶

Deprecations¶

Version 1.43.0 (2026-06-16)¶

New Features¶

• Registry: inference_engine_options["engine"] now accepts case-insensitive strings in addition to InferenceEngine enum members. Supported values are "vllm" and "python_generic" (for example, "vLLM", "VLLM", and "PYTHON_GENERIC" are all accepted). This applies to ModelVersion.create_service() and ModelVersion.run_batch().

from snowflake.ml.registry import Registry

registry = Registry(session)
mv = registry.get_model("my_model").version("v1")

mv.create_service(
    service_name="my_vllm_service",
    service_compute_pool="GPU_COMPUTE_POOL",
    gpu_requests="1",
    inference_engine_options={
        "engine": "vLLM",
        "engine_args_override": ["--max-model-len=4096"],
    },
)

job = mv.run_batch(
    compute_pool="GPU_COMPUTE_POOL",
    X=input_df,
    output_spec=output_spec,
    inference_engine_options={
        "engine": "python_generic",
    },
)

Bug Fixes¶

• Registry: Fixed explain() failing with type mismatch errors when passing a Snowpark DataFrame. Snowpark DataFrame columns are now explicitly cast to match model signature types before SQL generation. Table functions (used by explain) enforce stricter type coercion than scalar functions (used by predict), so columns that were implicitly coerced for predict would be rejected by explain.

Behavior Changes¶

• Registry: SentenceTransformers models no longer hard-code a default inference batch size of 32. When batch_size is not specified at log_model time and is not overridden at inference, the sentence-transformers library default is used. To pin a specific size, pass batch_size via log_model(..., options={"batch_size": N}) or as an inference-time parameter. All SentenceTransformers models logged with this release require a client/serving runtime on version 1.43.0 or later (regardless of whether batch_size was specified).

Deprecations¶

Version 1.42.0 (2026-06-11)¶

New Features¶

• Registry: Support truncate_dim for SentenceTransformers models. Truncation set at construction (SentenceTransformer(..., truncate_dim=N)) is captured when the model is logged and restored on reload. On clients with sentence-transformers 5.0.0 or later, truncate_dim is also exposed as a runtime parameter in the model signature.

• Registry: log_model() now captures a representative row from sample_input_data and stores it alongside the model so inference code snippets shown in the model registry UI can be pre-filled with realistic values. Pass options={"capture_sample_input_data": False} to opt out (e.g., for sensitive data); generic placeholder values will be used in the snippets instead.

Bug Fixes¶

• Registry: Pin transformers<5 for additional HuggingFace pipeline tasks removed in transformers 5.x (summarization, text2text-generation, question-answering, and translation_*), preventing deployment failures from missing pipeline classes.

Behavior Changes¶

Deprecations¶

Version 1.41.0 (2026-05-27)¶

New Features¶

• Registry: Extended ParamSpec support to MLflow PyFunc, LightGBM, XGBoost, and CatBoost models. Parameters declared in the model signature can now be passed at inference time.

Bug Fixes¶

Behavior Changes¶

Deprecations¶

Version 1.40.0 (2026-05-19)¶

New Features¶

Bug Fixes¶

• Feature Store: get_feature_view() now correctly preserves online_config for feature views whose names require SQL quoting (mixed case or special characters such as a space). Previously the returned FeatureView reported online=False even when online was enabled at registration, causing read_feature_view(store_type=ONLINE) to fail with “Online store is not enabled”.

Behavior Changes¶

• Registry: log_model(..., options=...) now rejects unknown top-level keys and unknown keys inside each method_options entry (for example misspellings or options that belong to a different model framework). Previously those keys were ignored. Validation is based on the public save-option TypedDicts and runs before dependency reconciliation.

• Registry: use_gpu is declared on the CatBoost, XGBoost, PyTorch, and TorchScript save-option TypedDicts to match the save paths that already read this flag.

Deprecations¶

Version 1.39.0¶

New Features¶

Bug Fixes¶

• Feature Store: update_feature_view now correctly handles every transition between duration-based and CRON-based refresh_freq. Previously a CRON expression was forwarded to the Dynamic Table’s TARGET_LAG (which Snowflake rejects), duration → cron failed because the companion Task did not yet exist, and cron → duration left the Task orphaned and continuing to fire on its old schedule. All four (old, new) transitions now correctly create, alter, or drop the Task as needed, with symmetric rollback on failure.

Behavior Changes¶

• Feature Store: For CRON-based feature views, get_feature_view(), list_feature_views(), and register_feature_view() now return the original cron expression as refresh_freq instead of "DOWNSTREAM". The underlying Dynamic Table still uses TARGET_LAG = 'DOWNSTREAM' with a companion Task — this is purely a display change so the round-trip preserves what the user passed in.

• Registry: When target_platforms includes WAREHOUSE and pip installs are needed without a user-supplied pip artifact repository, log_model injects snowflake.snowpark.pypi_shared_repository after verifying access. This applies to explicit pip_requirements and to pip-only packaging when there are no user conda dependencies. If pypi_shared_repository is inaccessible in the pip-only case, automatic packaging dependencies fall back to conda instead of forcing pip-only without an index.

Deprecations¶

Version 1.38.0¶

New Features¶

• Registry: Auto-inferred signatures for HuggingFace pipeline models now include task-specific ParamSpecs, allowing users to control inference behavior at prediction time via params. Generative tasks expose GenerationConfig parameters (e.g. temperature, max_new_tokens, top_p); non-generative tasks expose their own task-specific parameters (e.g. top_k for fill-mask, aggregation_strategy for zero-shot-classification).

Bug Fixes¶

Behavior Changes¶

• Registry: For HuggingFace text-generation pipelines whose tokenizer defines a chat template, the auto-inferred signature now matches the OpenAI Chat Completions API (_OPENAI_CHAT_SIGNATURE_WITH_PARAMS_SPEC). Inputs are a single messages column and inference controls (temperature, max_completion_tokens, stop, n, stream, top_p, frequency_penalty, presence_penalty) move from inputs to params with default values. Predictions return the OpenAI response shape (id, object, created, model, choices, usage); the generated text is at choices[0].message.content instead of outputs[0].generated_text.

• Registry: For HuggingFace image-text-to-text, video-text-to-text, and audio-text-to-text pipelines, the auto-inferred signature now uses _OPENAI_CHAT_SIGNATURE_WITH_PARAMS_SPEC instead of _OPENAI_CHAT_SIGNATURE_SPEC. The input column set narrows to just messages, and the inference controls move to params with default values; the output schema is unchanged.

Deprecations¶

Version 1.37.0 (2026-04-29)¶

New Features¶

• Registry (PrPr): The model_init_once save option now applies to TABLE_FUNCTION model methods (including partitioned table functions) for warehouse deployments, in addition to scalar FUNCTION methods. The generated handler uses the same eager @udf_init_once model load path as UDFs.

• Experiment Tracking: end_run now accepts an optional status argument ("FINISHED" or "FAILED") to explicitly set the final run status. When a run’s context manager exits with an exception, the status is automatically set to "FAILED".

Bug Fixes¶

• Registry: Fixed log_model() failing for some Snowpark ML Pipeline models with explainability when the full pipeline could not be converted to a native object; task inference now uses the final estimator instead.

• Registry: run_batch() now raises a clear ValueError when a partitioned model’s output signature includes the partition column. Previously this produced duplicate columns in the output, causing cryptic Ray/Arrow errors (AttributeError or KeyError) deep in the batch inference pipeline.

Behavior Changes¶

Deprecations¶

Version 1.36.0 (2026-04-22)¶

New Features¶

• Feature Store: for latency-sensitive online feature view reads, set use_session_warehouse=True to re-use the warehouse from the current session and achieve the best latency.

Bug Fixes¶

• Registry: Pin transformers<5 when saving HuggingFace pipeline models that use tasks removed in transformers 5.x (image-to-text, visual-question-answering, conversational), preventing deployment failures from missing pipeline classes.

Behavior Changes¶

• Registry: Logging a model with pip_requirements and no artifact_repository_map now raises a ValueError when target_platforms explicitly includes "WAREHOUSE". Previously this combination would silently proceed, resulting in the warehouse deployment being silently dropped.

Deprecations¶

Version 1.35.0 (2026-04-17)¶

New Features¶

Bug Fixes¶

• Registry: Fixed ParamSpec.from_mlflow_spec dropping shape, which caused shaped scalar params (e.g., array of ints) from MLflow to fail validation during model import.

Behavior Changes¶

• ML Jobs: spec_overrides now validates container keys and warns when keys other than name and secrets are provided. Additional keys are not officially supported and may not behave as expected.

Deprecations¶

Version 1.34.0 (2026-04-08)¶

New Features¶

• Experiment Tracking: Added list_params and list_metrics methods to retrieve parameters and metrics for runs within an experiment. Both methods return a Dataframe and accept an optional run_name argument to filter to a specific run.

Bug Fixes¶

• Feature Store: Fixed generate_dataset()/generate_training_set() SQL generation for Unicode and case-sensitive identifiers (for example Japanese column names), ensuring columns are quoted exactly once and preventing SQL compilation errors such as unexpected '<column_name>'.

Behavior Changes¶

• Registry: enable_explainability=True is now allowed for SPCS-only and non-warehouse-runnable models. Previously this raised a ValueError.

Deprecations¶

Version 1.33.0 (2026-03-31)¶

New Features¶

• Registry: Extended ParamSpec support to PyTorch models. Parameters declared in the model signature can now be passed at inference time, matching the existing support for custom models.
• Registry (PrPr): Added model_init_once model save option to log_model. When set to True, the model is loaded once per worker process at startup, eliminating model-loading overhead. Defaults to False.
• Experiment Tracking live logging is out of private preview and will become generally available over the next few weeks.
• Registry: Bumped transformers upper bound to <6, adding compatibility with Hugging Face Transformers v5.

Bug Fixes¶

• Experiment Tracking live logging: Fixed a bug where the tracebacks for uncaught exceptions were not logged when the code was running in a with exp.start_run() block.
• Registry: Fixed a thread leak in create_service(block=True) where the log-streaming thread could outlive the error handler when service deployment fails.

Behavior Changes¶

Deprecations¶

Version 1.32.0 (2026-03-26)¶

New Features¶

• Registry: Extended ParamSpec support to scikit-learn models. Parameters declared in the model signature can now be passed at inference time, matching the existing support for custom models.
• Registry: infer_signature() now accepts a dict for the params argument (e.g., params={"temperature": 0.7, "max_tokens": 100}), automatically inferring ParamSpec objects from the Python value types. The existing Sequence[BaseParamSpec] form is still supported.

Bug Fixes¶

• Experiment Tracking live logging (PrPr): Fixed a bug where the tracebacks for uncaught exceptions were not logged when the code was running inside Snowflake notebooks.

Behavior Changes¶

• Registry: The models from huggingface.TransformersPipeline and transformers.Pipeline will default to safetensors file format by default by using safe_serialization=True parameter to the save_pretrained call. This creates safetensors files instead of PyTorch binaries or pt files, addressing security vulnerability CVE-2025-32434.

Deprecations¶

Version 1.31.0 (2026-03-19)¶

New Features¶

Bug Fixes¶

• Registry: Fixed custom model handler to explicitly persist PARTITIONED=False in model metadata for @inference_api methods. Previously, custom models using @inference_api with TABLE_FUNCTION type were incorrectly reported as partitioned.
• Registry: Fixed sklearn and SnowML model handlers to explicitly persist PARTITIONED=False metadata for the explain method, preventing it from being incorrectly reported as partitioned at read time.
• Registry: Fixed inconsistent infer_signature for columns containing NaN values. Previously, the inferred dtype (DOUBLE vs INT64) depended on whether NaN rows survived truncation, causing downstream validation failures. Now, the original pandas column dtype is respected during infer_signature, so columns with mix of integer and np.nan or None are consistently inferred as DOUBLE regardless of NaN position.
• Registry: Fixed a bug where invalid parameter types (e.g., passing a string to an integer parameter) were silently coerced instead of raising an error. Parameters are now strictly validated against their declared types before inference. This may break code that relied on the silent coercion behavior (e.g., mv.run(..., params={"max_tokens": "100"}) or mv.run(..., params={"some_int_param": True})).
• ML Job: Fixed explicitly named MLJobDefinition objects so they can be overwritten in place and invoked repeatedly without job name collisions; the generate_suffix argument has been removed.
• Experiment Tracking live logging (PrPr): Fixed a bug that generated too many lines of log that consist entirely of whitespace.

Behavior Changes¶

• Registry: Warehouse partitioned model inference no longer includes non-partition input columns (which were always NULL) in the output. The output now contains only the model’s output columns and the partition column.

Deprecations¶

Version 1.30.0¶

New Features¶

• Experiment Tracking live logging (PrPr): In SPCS, call set_live_logging_status(True) to automatically capture and persist outputs to stdout and stderr while a run is active. The captured logs can be viewed from the Experiments UI.
• Registry: Support logging MLflow models created via mlflow.*.save_model() to the Snowflake Model Registry. Previously, only models logged through mlflow.*.log_model() were supported. This also enables logging custom mlflow.pyfunc.PythonModel subclasses saved locally.

Bug Fixes¶

• Registry: Fixed Prophet model handler to correctly mark the predict method as partitioned, ensuring it uses a partitioned TABLE_FUNCTION when deployed to Snowflake.
• Registry: Support text-generation models without chat template. The model will have signatures to automatically take plain strings as input without needing to specify the signatures in log_model. The signature can be overridden if the user chooses to.

Behavior Changes¶

• Registry: Huggingface models with task text-generation that do not have chat templates will be logged with signature that supports plain text (string) as input.

Deprecations¶

• Registry: Removed support for logging Hugging Face Pipelines in config-only mode. Config-only models could not run in warehouse and required an External Access Integration (EAI) with egress to Hugging Face hosts. Use remote logging or local download instead — these approaches support warehouse execution and store model weights at log time, enabling fully air-gapped services.

Version 1.29.0 (2026-02-24)¶

New Features¶

• Model serving: Introducing InferenceEngine.PYTHON_GENERIC enum value. Users can pass InferenceEngine.PYTHON_GENERIC to use a Python-based inference server for model serving.

Bug Fixes¶

• Registry: Fixed a bug where using inference parameters (ParamSpec) with table function or partitioned model methods would fail at runtime with a NameError.
• Registry: Fixed a bug where MLflow models with string columns failed during inference with “Can not safely convert string to <U0>” errors due to MLflow’s schema validation not handling pd.StringDtype correctly.

Behavior Changes¶

• Dependencies: Removed snowflake.core as a runtime dependency. If you use snowflake.core (e.g., TaskDag, Root, Task), add snowflake.core explicitly to your dependencies.

Deprecations¶

Version 1.28.0 (2026-02-17)¶

New Features¶

Bug Fixes¶

Behavior Changes¶

Deprecations¶

Version 1.27.0 (2026-02-12)¶

Bug fixes¶

Model Registry bug fixes:

• Fixed failure of model_version.run caused by requiring READ privilege on the model instead of USAGE when the user’s role had only the USAGE privilege.

Feature store bug fixes:

• Fixed failure of register_feature_view with overwrite=True when the existing feature view is external and the new feature view is managed, or vice versa.

Version 1.26.0 (2026-02-05)¶

New features¶

New Model Registry features:

• Model signatures can now include inference parameters via ParamSpec, allowing you to define constant parameters to be passed at inference time without including them in the input data. Example:
  Copy code

  import pandas as pd
  from snowflake.ml.model import custom_model, model_signature
  from snowflake.ml.registry import Registry
  
  # Define a custom model with inference parameters
  class MyModelWithParams(custom_model.CustomModel):
   @custom_model.inference_api
   def predict(
       self,
       input_df: pd.DataFrame,
       *,
       temperature: float = 1.0,  # keyword-only param with default
   ) -> pd.DataFrame:
       return pd.DataFrame({"output": input_df["feature"] * temperature})
  
  # Create sample data
  model = MyModelWithParams(custom_model.ModelContext())
  sample_input = pd.DataFrame({"feature": [1.0, 2.0, 3.0]})
  sample_output = model.predict(sample_input, temperature=1.0)
  
  # Define ParamSpec for the inference parameter
  params = [
   model_signature.ParamSpec(
       name="temperature",
       dtype=model_signature.DataType.FLOAT,
       default_value=1.0,
   ),
  ]
  
  # Infer signature with params
  sig = model_signature.infer_signature(
   input_data=sample_input,
   output_data=sample_output,
   params=params,
  )
  
  # Log model with the signature
  registry = Registry(session)
  mv = registry.log_model(
   model=model,
   model_name="my_model_with_params",
   version_name="v1",
   signatures={"predict": sig},
  )
  
  # Run inference with custom parameter value
  result = mv.run(sample_input, function_name="predict", params={"temperature": 2.0})
  

New Feature Store features:

• New auto_prefix parameter and with_name method to prevent column name collisions when joining multiple feature views in dataset generation.
• Dynamic Iceberg tables can now be used as backing storage for Feature Views. Use StorageConfig with StorageFormat.ICEBERG to store data in Apache Iceberg format on external cloud storage. A new default_iceberg_external_volume parameter is available in FeatureStore to set a default external volume for Iceberg feature views.

Version 1.25.1 (2026-02-03)¶

No public-facing changes. This release includes changes to a preview feature that has not been publicly announced.

Version 1.25.0 (2026-01-28)¶

New features¶

New Model Serving features:

• The create_service method accepts a new autocapture argument to indicate whether inference data should be captured (see Autocapture inference logs for realtime inference).
• The create_service and log_model_and_create_service methods now accept an optional min_instances argument to specify the minimum number of instances for the service. The service automatically scales between the specified minimum and maximum instances based on traffic and hardware utilization. If min_instances is 0, the service automatically suspends when no traffic is detected for a period of time. The default value for min_instances is 0.

Version 1.24.0 (2026-01-22)¶

New features¶

New Feature Store features:

• Tile-based aggregation support using a new Feature API for efficient and point-in-time correct time-series feature computation using pre-computed tiles.

New Model Registry features:

• SentenceTransformer models now support automatic signature inference. When logging a SentenceTransformer model, sample_input_data is optional. The signature is automatically inferred from the model’s embedding dimension when sample input data is not provided.. The encode, encode_query, encode_document, encode_queries, encode_documents methods are supported.
  Copy code

  import sentence_transformers
  from snowflake.ml.registry import Registry
  
  # Create model
  model = sentence_transformers.SentenceTransformer("all-MiniLM-L6-v2")
  
  # Log model without sample_input_data - signature is auto-inferred
  registry = Registry(session)
  mv = registry.log_model(
   model=model,
   model_name="my_sentence_transformer",
   version_name="v1",
  )
  
  # Run inference with auto-inferred signature (input: "text", output: "output")
  import pandas as pd
  result = mv.run(pd.DataFrame({"text": ["Hello world"]}))
  

Version 1.23.0 (2026-01-165)¶

New features¶

New ML Jobs features:

• ML Jobs now support Python 3.11 and Python 3.12. Jobs automatically select a runtime environment matching the client Python version.

Bug fixes¶

Model Registry bug fixes:

• Empty output in HuggingFace’s Token Classification (Named Entity Recognition) models no longer causes failures.

Model Serving bug fixes:

• Container statuses are now correctly reported and should not be blank.

Version 1.22.0 (2026-01-09)¶

New features¶

New Model Registry features:

• You can now remotely log a transformer pipeline model using a Snowpark Container Services (SPCS) job.
  Copy code

  # create reference to the model
  model = huggingface.TransformersPipeline(
   model="TinyLlama/TinyLlama-1.1B-Chat-v1.0",
   task="text-generation",
  )
  
  # Remotely log the model, a SPCS job will run async and log the model
  mv = registry.log_model(
   model=model,
   model_name="tinyllama_remote_log",
   target_platforms=["SNOWPARK_CONTAINER_SERVICES"],
   signatures=openai_signatures.OPENAI_CHAT_SIGNATURE,
  )
  

Version 1.21.0 (2026-01-05)¶

Behavior changes¶

ML Jobs behavior changes:

• The behavior of the additional_payloads parameter is changing. Use the imports argument to declare additional dependencies, such as ZIP files and Python modules. Local directories and Python files are automatically compressed, and their internal layout is determined by the specified import path. The import path applies only to local directories, Python files, and staged python files; it has no effect on other import types. When referencing files in a stage, only individual files are supported, not directories.

Experiment Tracking behavior changes:

• ExperimentTracking is now a singleton class.

Bug fixes¶

Experiment Tracking bug fixes:

• Reaching the run metadata size limit in log_metrics or log_params now issues a warning instead of raising an exception.

Model Registry bug fixes:

• ModelVersion.run now raises a ValueError if the model is a SPCS-only model and service_name is not provided.

New preview features¶

• The create_service method now accepts the Boolean argument autocapture to indicate whether inference data is automatically captured.

New release features¶

New Model Registry features:

• The new snowflake.ml.model.models.huggingface.TransformersPipeline class is intended to replace snowflake.ml.model.models.huggingface_pipeline.HuggingfacePipelineModel, although the older class is not yet deprecated. The new class knows model signatures for common tasks so that you do not need to specify them manually. The supported tasks are currently:

  • fill-mask
  • question-answering
  • summarization
  • table-question-answering
  • text2text-generation
  • text-classification (alias sentiment-analysis)
  • text-generation
  • token-classification (alias ner)
  • translation
  • translation_xx_to_yy
  • zero-shot-classification (lets you log models without loading them into memory)

• The list_services API now shows an internal endpoint that can be called from another SPCS node or notebook without Enterprise Application Integration. It also indicates whether autocapture is enabled for each service.

New DataConnector features:

• New to_huggingface_dataset method converts Snowflake data to HuggingFace datasets. Supports both in-memory Dataset (streaming=False) and streaming IterableDataset (streaming=True) modes.

Deprecation notices¶

• The snowflake.ml.model.models.huggingface_pipeline.HuggingfacePipelineModel class has been deprecated and will be removed in a future release.