yandexcloud 0.326.0__py3-none-any.whl → 0.328.0__py3-none-any.whl

yandex/cloud/ai/tuning/v1/tuning_service_pb2.pyi

@@ -0,0 +1,585 @@
+"""
+@generated by mypy-protobuf.  Do not edit manually!
+isort:skip_file
+"""
+
+import builtins
+import collections.abc
+import google.protobuf.descriptor
+import google.protobuf.internal.containers
+import google.protobuf.message
+import typing
+import yandex.cloud.ai.tuning.v1.tuning_optimizers_pb2
+import yandex.cloud.ai.tuning.v1.tuning_schedulers_pb2
+import yandex.cloud.ai.tuning.v1.tuning_task_pb2
+import yandex.cloud.ai.tuning.v1.tuning_types_pb2
+
+DESCRIPTOR: google.protobuf.descriptor.FileDescriptor
+
+@typing.final
+class ListTuningsRequest(google.protobuf.message.Message):
+    DESCRIPTOR: google.protobuf.descriptor.Descriptor
+
+    FOLDER_ID_FIELD_NUMBER: builtins.int
+    PAGE_SIZE_FIELD_NUMBER: builtins.int
+    PAGE_TOKEN_FIELD_NUMBER: builtins.int
+    folder_id: builtins.str
+    """Required field. ID of the folder to list tunings in."""
+    page_size: builtins.int
+    """Maximum number of tuning tasks to return per page."""
+    page_token: builtins.str
+    """Token to retrieve the next page of results."""
+    def __init__(
+        self,
+        *,
+        folder_id: builtins.str = ...,
+        page_size: builtins.int = ...,
+        page_token: builtins.str = ...,
+    ) -> None: ...
+    def ClearField(self, field_name: typing.Literal["folder_id", b"folder_id", "page_size", b"page_size", "page_token", b"page_token"]) -> None: ...
+
+global___ListTuningsRequest = ListTuningsRequest
+
+@typing.final
+class ListTuningsResponse(google.protobuf.message.Message):
+    DESCRIPTOR: google.protobuf.descriptor.Descriptor
+
+    TUNING_TASKS_FIELD_NUMBER: builtins.int
+    NEXT_PAGE_TOKEN_FIELD_NUMBER: builtins.int
+    next_page_token: builtins.str
+    """Token to retrieve the next page of results."""
+    @property
+    def tuning_tasks(self) -> google.protobuf.internal.containers.RepeatedCompositeFieldContainer[yandex.cloud.ai.tuning.v1.tuning_task_pb2.TuningTask]:
+        """List of tuning tasks in the specified folder."""
+
+    def __init__(
+        self,
+        *,
+        tuning_tasks: collections.abc.Iterable[yandex.cloud.ai.tuning.v1.tuning_task_pb2.TuningTask] | None = ...,
+        next_page_token: builtins.str = ...,
+    ) -> None: ...
+    def ClearField(self, field_name: typing.Literal["next_page_token", b"next_page_token", "tuning_tasks", b"tuning_tasks"]) -> None: ...
+
+global___ListTuningsResponse = ListTuningsResponse
+
+@typing.final
+class DescribeTuningRequest(google.protobuf.message.Message):
+    DESCRIPTOR: google.protobuf.descriptor.Descriptor
+
+    TUNING_TASK_ID_FIELD_NUMBER: builtins.int
+    tuning_task_id: builtins.str
+    def __init__(
+        self,
+        *,
+        tuning_task_id: builtins.str = ...,
+    ) -> None: ...
+    def ClearField(self, field_name: typing.Literal["tuning_task_id", b"tuning_task_id"]) -> None: ...
+
+global___DescribeTuningRequest = DescribeTuningRequest
+
+@typing.final
+class DescribeTuningResponse(google.protobuf.message.Message):
+    DESCRIPTOR: google.protobuf.descriptor.Descriptor
+
+    TUNING_TASK_FIELD_NUMBER: builtins.int
+    @property
+    def tuning_task(self) -> yandex.cloud.ai.tuning.v1.tuning_task_pb2.TuningTask: ...
+    def __init__(
+        self,
+        *,
+        tuning_task: yandex.cloud.ai.tuning.v1.tuning_task_pb2.TuningTask | None = ...,
+    ) -> None: ...
+    def HasField(self, field_name: typing.Literal["tuning_task", b"tuning_task"]) -> builtins.bool: ...
+    def ClearField(self, field_name: typing.Literal["tuning_task", b"tuning_task"]) -> None: ...
+
+global___DescribeTuningResponse = DescribeTuningResponse
+
+@typing.final
+class CancelTuningRequest(google.protobuf.message.Message):
+    DESCRIPTOR: google.protobuf.descriptor.Descriptor
+
+    TUNING_TASK_ID_FIELD_NUMBER: builtins.int
+    tuning_task_id: builtins.str
+    def __init__(
+        self,
+        *,
+        tuning_task_id: builtins.str = ...,
+    ) -> None: ...
+    def ClearField(self, field_name: typing.Literal["tuning_task_id", b"tuning_task_id"]) -> None: ...
+
+global___CancelTuningRequest = CancelTuningRequest
+
+@typing.final
+class CancelTuningResponse(google.protobuf.message.Message):
+    DESCRIPTOR: google.protobuf.descriptor.Descriptor
+
+    TUNING_TASK_ID_FIELD_NUMBER: builtins.int
+    tuning_task_id: builtins.str
+    def __init__(
+        self,
+        *,
+        tuning_task_id: builtins.str = ...,
+    ) -> None: ...
+    def ClearField(self, field_name: typing.Literal["tuning_task_id", b"tuning_task_id"]) -> None: ...
+
+global___CancelTuningResponse = CancelTuningResponse
+
+@typing.final
+class TuningResponse(google.protobuf.message.Message):
+    DESCRIPTOR: google.protobuf.descriptor.Descriptor
+
+    TUNING_TASK_ID_FIELD_NUMBER: builtins.int
+    STATUS_FIELD_NUMBER: builtins.int
+    TARGET_MODEL_URI_FIELD_NUMBER: builtins.int
+    tuning_task_id: builtins.str
+    status: yandex.cloud.ai.tuning.v1.tuning_task_pb2.TuningTask.Status.ValueType
+    target_model_uri: builtins.str
+    def __init__(
+        self,
+        *,
+        tuning_task_id: builtins.str = ...,
+        status: yandex.cloud.ai.tuning.v1.tuning_task_pb2.TuningTask.Status.ValueType = ...,
+        target_model_uri: builtins.str = ...,
+    ) -> None: ...
+    def ClearField(self, field_name: typing.Literal["status", b"status", "target_model_uri", b"target_model_uri", "tuning_task_id", b"tuning_task_id"]) -> None: ...
+
+global___TuningResponse = TuningResponse
+
+@typing.final
+class TuningMetadata(google.protobuf.message.Message):
+    DESCRIPTOR: google.protobuf.descriptor.Descriptor
+
+    TUNING_TASK_ID_FIELD_NUMBER: builtins.int
+    STATUS_FIELD_NUMBER: builtins.int
+    tuning_task_id: builtins.str
+    status: yandex.cloud.ai.tuning.v1.tuning_task_pb2.TuningTask.Status.ValueType
+    def __init__(
+        self,
+        *,
+        tuning_task_id: builtins.str = ...,
+        status: yandex.cloud.ai.tuning.v1.tuning_task_pb2.TuningTask.Status.ValueType = ...,
+    ) -> None: ...
+    def ClearField(self, field_name: typing.Literal["status", b"status", "tuning_task_id", b"tuning_task_id"]) -> None: ...
+
+global___TuningMetadata = TuningMetadata
+
+@typing.final
+class TuningRequest(google.protobuf.message.Message):
+    DESCRIPTOR: google.protobuf.descriptor.Descriptor
+
+    @typing.final
+    class LabelsEntry(google.protobuf.message.Message):
+        DESCRIPTOR: google.protobuf.descriptor.Descriptor
+
+        KEY_FIELD_NUMBER: builtins.int
+        VALUE_FIELD_NUMBER: builtins.int
+        key: builtins.str
+        value: builtins.str
+        def __init__(
+            self,
+            *,
+            key: builtins.str = ...,
+            value: builtins.str = ...,
+        ) -> None: ...
+        def ClearField(self, field_name: typing.Literal["key", b"key", "value", b"value"]) -> None: ...
+
+    @typing.final
+    class WeightedDataset(google.protobuf.message.Message):
+        DESCRIPTOR: google.protobuf.descriptor.Descriptor
+
+        DATASET_ID_FIELD_NUMBER: builtins.int
+        WEIGHT_FIELD_NUMBER: builtins.int
+        dataset_id: builtins.str
+        weight: builtins.float
+        def __init__(
+            self,
+            *,
+            dataset_id: builtins.str = ...,
+            weight: builtins.float = ...,
+        ) -> None: ...
+        def ClearField(self, field_name: typing.Literal["dataset_id", b"dataset_id", "weight", b"weight"]) -> None: ...
+
+    BASE_MODEL_URI_FIELD_NUMBER: builtins.int
+    TRAIN_DATASETS_FIELD_NUMBER: builtins.int
+    VALIDATION_DATASETS_FIELD_NUMBER: builtins.int
+    TEXT_TO_TEXT_COMPLETION_FIELD_NUMBER: builtins.int
+    TEXT_CLASSIFICATION_MULTILABEL_FIELD_NUMBER: builtins.int
+    TEXT_CLASSIFICATION_MULTICLASS_FIELD_NUMBER: builtins.int
+    NAME_FIELD_NUMBER: builtins.int
+    DESCRIPTION_FIELD_NUMBER: builtins.int
+    LABELS_FIELD_NUMBER: builtins.int
+    base_model_uri: builtins.str
+    """Format like a gpt://{folder_id}/yandex-gpt/latest"""
+    name: builtins.str
+    """common params"""
+    description: builtins.str
+    @property
+    def train_datasets(self) -> google.protobuf.internal.containers.RepeatedCompositeFieldContainer[global___TuningRequest.WeightedDataset]: ...
+    @property
+    def validation_datasets(self) -> google.protobuf.internal.containers.RepeatedCompositeFieldContainer[global___TuningRequest.WeightedDataset]: ...
+    @property
+    def text_to_text_completion(self) -> global___TextToTextCompletionTuningParams: ...
+    @property
+    def text_classification_multilabel(self) -> global___TextClassificationMultilabelParams: ...
+    @property
+    def text_classification_multiclass(self) -> global___TextClassificationMulticlassParams: ...
+    @property
+    def labels(self) -> google.protobuf.internal.containers.ScalarMap[builtins.str, builtins.str]: ...
+    def __init__(
+        self,
+        *,
+        base_model_uri: builtins.str = ...,
+        train_datasets: collections.abc.Iterable[global___TuningRequest.WeightedDataset] | None = ...,
+        validation_datasets: collections.abc.Iterable[global___TuningRequest.WeightedDataset] | None = ...,
+        text_to_text_completion: global___TextToTextCompletionTuningParams | None = ...,
+        text_classification_multilabel: global___TextClassificationMultilabelParams | None = ...,
+        text_classification_multiclass: global___TextClassificationMulticlassParams | None = ...,
+        name: builtins.str = ...,
+        description: builtins.str = ...,
+        labels: collections.abc.Mapping[builtins.str, builtins.str] | None = ...,
+    ) -> None: ...
+    def HasField(self, field_name: typing.Literal["text_classification_multiclass", b"text_classification_multiclass", "text_classification_multilabel", b"text_classification_multilabel", "text_to_text_completion", b"text_to_text_completion", "tuning_params", b"tuning_params"]) -> builtins.bool: ...
+    def ClearField(self, field_name: typing.Literal["base_model_uri", b"base_model_uri", "description", b"description", "labels", b"labels", "name", b"name", "text_classification_multiclass", b"text_classification_multiclass", "text_classification_multilabel", b"text_classification_multilabel", "text_to_text_completion", b"text_to_text_completion", "train_datasets", b"train_datasets", "tuning_params", b"tuning_params", "validation_datasets", b"validation_datasets"]) -> None: ...
+    def WhichOneof(self, oneof_group: typing.Literal["tuning_params", b"tuning_params"]) -> typing.Literal["text_to_text_completion", "text_classification_multilabel", "text_classification_multiclass"] | None: ...
+
+global___TuningRequest = TuningRequest
+
+@typing.final
+class TextToTextCompletionTuningParams(google.protobuf.message.Message):
+    DESCRIPTOR: google.protobuf.descriptor.Descriptor
+
+    @typing.final
+    class Scheduler(google.protobuf.message.Message):
+        DESCRIPTOR: google.protobuf.descriptor.Descriptor
+
+        LINEAR_FIELD_NUMBER: builtins.int
+        CONSTANT_FIELD_NUMBER: builtins.int
+        COSINE_FIELD_NUMBER: builtins.int
+        WARMUP_RATIO_FIELD_NUMBER: builtins.int
+        warmup_ratio: builtins.float
+        @property
+        def linear(self) -> yandex.cloud.ai.tuning.v1.tuning_schedulers_pb2.SchedulerLinear: ...
+        @property
+        def constant(self) -> yandex.cloud.ai.tuning.v1.tuning_schedulers_pb2.SchedulerConstant: ...
+        @property
+        def cosine(self) -> yandex.cloud.ai.tuning.v1.tuning_schedulers_pb2.SchedulerCosine: ...
+        def __init__(
+            self,
+            *,
+            linear: yandex.cloud.ai.tuning.v1.tuning_schedulers_pb2.SchedulerLinear | None = ...,
+            constant: yandex.cloud.ai.tuning.v1.tuning_schedulers_pb2.SchedulerConstant | None = ...,
+            cosine: yandex.cloud.ai.tuning.v1.tuning_schedulers_pb2.SchedulerCosine | None = ...,
+            warmup_ratio: builtins.float = ...,
+        ) -> None: ...
+        def HasField(self, field_name: typing.Literal["constant", b"constant", "cosine", b"cosine", "linear", b"linear", "type", b"type"]) -> builtins.bool: ...
+        def ClearField(self, field_name: typing.Literal["constant", b"constant", "cosine", b"cosine", "linear", b"linear", "type", b"type", "warmup_ratio", b"warmup_ratio"]) -> None: ...
+        def WhichOneof(self, oneof_group: typing.Literal["type", b"type"]) -> typing.Literal["linear", "constant", "cosine"] | None: ...
+
+    @typing.final
+    class Optimizer(google.protobuf.message.Message):
+        DESCRIPTOR: google.protobuf.descriptor.Descriptor
+
+        ADAMW_FIELD_NUMBER: builtins.int
+        @property
+        def adamw(self) -> yandex.cloud.ai.tuning.v1.tuning_optimizers_pb2.OptimizerAdamw: ...
+        def __init__(
+            self,
+            *,
+            adamw: yandex.cloud.ai.tuning.v1.tuning_optimizers_pb2.OptimizerAdamw | None = ...,
+        ) -> None: ...
+        def HasField(self, field_name: typing.Literal["adamw", b"adamw", "type", b"type"]) -> builtins.bool: ...
+        def ClearField(self, field_name: typing.Literal["adamw", b"adamw", "type", b"type"]) -> None: ...
+        def WhichOneof(self, oneof_group: typing.Literal["type", b"type"]) -> typing.Literal["adamw"] | None: ...
+
+    SEED_FIELD_NUMBER: builtins.int
+    LR_FIELD_NUMBER: builtins.int
+    N_SAMPLES_FIELD_NUMBER: builtins.int
+    ADDITIONAL_ARGUMENTS_FIELD_NUMBER: builtins.int
+    LORA_FIELD_NUMBER: builtins.int
+    PROMPT_TUNE_FIELD_NUMBER: builtins.int
+    SCHEDULER_FIELD_NUMBER: builtins.int
+    OPTIMIZER_FIELD_NUMBER: builtins.int
+    seed: builtins.int
+    lr: builtins.float
+    n_samples: builtins.int
+    additional_arguments: builtins.str
+    @property
+    def lora(self) -> yandex.cloud.ai.tuning.v1.tuning_types_pb2.TuningTypeLora: ...
+    @property
+    def prompt_tune(self) -> yandex.cloud.ai.tuning.v1.tuning_types_pb2.TuningTypePromptTune: ...
+    @property
+    def scheduler(self) -> global___TextToTextCompletionTuningParams.Scheduler: ...
+    @property
+    def optimizer(self) -> global___TextToTextCompletionTuningParams.Optimizer: ...
+    def __init__(
+        self,
+        *,
+        seed: builtins.int = ...,
+        lr: builtins.float = ...,
+        n_samples: builtins.int = ...,
+        additional_arguments: builtins.str = ...,
+        lora: yandex.cloud.ai.tuning.v1.tuning_types_pb2.TuningTypeLora | None = ...,
+        prompt_tune: yandex.cloud.ai.tuning.v1.tuning_types_pb2.TuningTypePromptTune | None = ...,
+        scheduler: global___TextToTextCompletionTuningParams.Scheduler | None = ...,
+        optimizer: global___TextToTextCompletionTuningParams.Optimizer | None = ...,
+    ) -> None: ...
+    def HasField(self, field_name: typing.Literal["lora", b"lora", "optimizer", b"optimizer", "prompt_tune", b"prompt_tune", "scheduler", b"scheduler", "tuning_type", b"tuning_type"]) -> builtins.bool: ...
+    def ClearField(self, field_name: typing.Literal["additional_arguments", b"additional_arguments", "lora", b"lora", "lr", b"lr", "n_samples", b"n_samples", "optimizer", b"optimizer", "prompt_tune", b"prompt_tune", "scheduler", b"scheduler", "seed", b"seed", "tuning_type", b"tuning_type"]) -> None: ...
+    def WhichOneof(self, oneof_group: typing.Literal["tuning_type", b"tuning_type"]) -> typing.Literal["lora", "prompt_tune"] | None: ...
+
+global___TextToTextCompletionTuningParams = TextToTextCompletionTuningParams
+
+@typing.final
+class TextClassificationMultilabelParams(google.protobuf.message.Message):
+    DESCRIPTOR: google.protobuf.descriptor.Descriptor
+
+    @typing.final
+    class Scheduler(google.protobuf.message.Message):
+        DESCRIPTOR: google.protobuf.descriptor.Descriptor
+
+        LINEAR_FIELD_NUMBER: builtins.int
+        CONSTANT_FIELD_NUMBER: builtins.int
+        COSINE_FIELD_NUMBER: builtins.int
+        WARMUP_RATIO_FIELD_NUMBER: builtins.int
+        warmup_ratio: builtins.float
+        @property
+        def linear(self) -> yandex.cloud.ai.tuning.v1.tuning_schedulers_pb2.SchedulerLinear: ...
+        @property
+        def constant(self) -> yandex.cloud.ai.tuning.v1.tuning_schedulers_pb2.SchedulerConstant: ...
+        @property
+        def cosine(self) -> yandex.cloud.ai.tuning.v1.tuning_schedulers_pb2.SchedulerCosine: ...
+        def __init__(
+            self,
+            *,
+            linear: yandex.cloud.ai.tuning.v1.tuning_schedulers_pb2.SchedulerLinear | None = ...,
+            constant: yandex.cloud.ai.tuning.v1.tuning_schedulers_pb2.SchedulerConstant | None = ...,
+            cosine: yandex.cloud.ai.tuning.v1.tuning_schedulers_pb2.SchedulerCosine | None = ...,
+            warmup_ratio: builtins.float = ...,
+        ) -> None: ...
+        def HasField(self, field_name: typing.Literal["constant", b"constant", "cosine", b"cosine", "linear", b"linear", "type", b"type"]) -> builtins.bool: ...
+        def ClearField(self, field_name: typing.Literal["constant", b"constant", "cosine", b"cosine", "linear", b"linear", "type", b"type", "warmup_ratio", b"warmup_ratio"]) -> None: ...
+        def WhichOneof(self, oneof_group: typing.Literal["type", b"type"]) -> typing.Literal["linear", "constant", "cosine"] | None: ...
+
+    @typing.final
+    class Optimizer(google.protobuf.message.Message):
+        DESCRIPTOR: google.protobuf.descriptor.Descriptor
+
+        ADAMW_FIELD_NUMBER: builtins.int
+        @property
+        def adamw(self) -> yandex.cloud.ai.tuning.v1.tuning_optimizers_pb2.OptimizerAdamw: ...
+        def __init__(
+            self,
+            *,
+            adamw: yandex.cloud.ai.tuning.v1.tuning_optimizers_pb2.OptimizerAdamw | None = ...,
+        ) -> None: ...
+        def HasField(self, field_name: typing.Literal["adamw", b"adamw", "type", b"type"]) -> builtins.bool: ...
+        def ClearField(self, field_name: typing.Literal["adamw", b"adamw", "type", b"type"]) -> None: ...
+        def WhichOneof(self, oneof_group: typing.Literal["type", b"type"]) -> typing.Literal["adamw"] | None: ...
+
+    SEED_FIELD_NUMBER: builtins.int
+    LR_FIELD_NUMBER: builtins.int
+    N_SAMPLES_FIELD_NUMBER: builtins.int
+    ADDITIONAL_ARGUMENTS_FIELD_NUMBER: builtins.int
+    LORA_FIELD_NUMBER: builtins.int
+    PROMPT_TUNE_FIELD_NUMBER: builtins.int
+    SCHEDULER_FIELD_NUMBER: builtins.int
+    OPTIMIZER_FIELD_NUMBER: builtins.int
+    seed: builtins.int
+    lr: builtins.float
+    n_samples: builtins.int
+    additional_arguments: builtins.str
+    @property
+    def lora(self) -> yandex.cloud.ai.tuning.v1.tuning_types_pb2.TuningTypeLora: ...
+    @property
+    def prompt_tune(self) -> yandex.cloud.ai.tuning.v1.tuning_types_pb2.TuningTypePromptTune: ...
+    @property
+    def scheduler(self) -> global___TextClassificationMultilabelParams.Scheduler: ...
+    @property
+    def optimizer(self) -> global___TextClassificationMultilabelParams.Optimizer: ...
+    def __init__(
+        self,
+        *,
+        seed: builtins.int = ...,
+        lr: builtins.float = ...,
+        n_samples: builtins.int = ...,
+        additional_arguments: builtins.str = ...,
+        lora: yandex.cloud.ai.tuning.v1.tuning_types_pb2.TuningTypeLora | None = ...,
+        prompt_tune: yandex.cloud.ai.tuning.v1.tuning_types_pb2.TuningTypePromptTune | None = ...,
+        scheduler: global___TextClassificationMultilabelParams.Scheduler | None = ...,
+        optimizer: global___TextClassificationMultilabelParams.Optimizer | None = ...,
+    ) -> None: ...
+    def HasField(self, field_name: typing.Literal["lora", b"lora", "optimizer", b"optimizer", "prompt_tune", b"prompt_tune", "scheduler", b"scheduler", "tuning_type", b"tuning_type"]) -> builtins.bool: ...
+    def ClearField(self, field_name: typing.Literal["additional_arguments", b"additional_arguments", "lora", b"lora", "lr", b"lr", "n_samples", b"n_samples", "optimizer", b"optimizer", "prompt_tune", b"prompt_tune", "scheduler", b"scheduler", "seed", b"seed", "tuning_type", b"tuning_type"]) -> None: ...
+    def WhichOneof(self, oneof_group: typing.Literal["tuning_type", b"tuning_type"]) -> typing.Literal["lora", "prompt_tune"] | None: ...
+
+global___TextClassificationMultilabelParams = TextClassificationMultilabelParams
+
+@typing.final
+class TextClassificationMulticlassParams(google.protobuf.message.Message):
+    DESCRIPTOR: google.protobuf.descriptor.Descriptor
+
+    @typing.final
+    class Scheduler(google.protobuf.message.Message):
+        DESCRIPTOR: google.protobuf.descriptor.Descriptor
+
+        LINEAR_FIELD_NUMBER: builtins.int
+        CONSTANT_FIELD_NUMBER: builtins.int
+        COSINE_FIELD_NUMBER: builtins.int
+        WARMUP_RATIO_FIELD_NUMBER: builtins.int
+        warmup_ratio: builtins.float
+        @property
+        def linear(self) -> yandex.cloud.ai.tuning.v1.tuning_schedulers_pb2.SchedulerLinear: ...
+        @property
+        def constant(self) -> yandex.cloud.ai.tuning.v1.tuning_schedulers_pb2.SchedulerConstant: ...
+        @property
+        def cosine(self) -> yandex.cloud.ai.tuning.v1.tuning_schedulers_pb2.SchedulerCosine: ...
+        def __init__(
+            self,
+            *,
+            linear: yandex.cloud.ai.tuning.v1.tuning_schedulers_pb2.SchedulerLinear | None = ...,
+            constant: yandex.cloud.ai.tuning.v1.tuning_schedulers_pb2.SchedulerConstant | None = ...,
+            cosine: yandex.cloud.ai.tuning.v1.tuning_schedulers_pb2.SchedulerCosine | None = ...,
+            warmup_ratio: builtins.float = ...,
+        ) -> None: ...
+        def HasField(self, field_name: typing.Literal["constant", b"constant", "cosine", b"cosine", "linear", b"linear", "type", b"type"]) -> builtins.bool: ...
+        def ClearField(self, field_name: typing.Literal["constant", b"constant", "cosine", b"cosine", "linear", b"linear", "type", b"type", "warmup_ratio", b"warmup_ratio"]) -> None: ...
+        def WhichOneof(self, oneof_group: typing.Literal["type", b"type"]) -> typing.Literal["linear", "constant", "cosine"] | None: ...
+
+    @typing.final
+    class Optimizer(google.protobuf.message.Message):
+        DESCRIPTOR: google.protobuf.descriptor.Descriptor
+
+        ADAMW_FIELD_NUMBER: builtins.int
+        @property
+        def adamw(self) -> yandex.cloud.ai.tuning.v1.tuning_optimizers_pb2.OptimizerAdamw: ...
+        def __init__(
+            self,
+            *,
+            adamw: yandex.cloud.ai.tuning.v1.tuning_optimizers_pb2.OptimizerAdamw | None = ...,
+        ) -> None: ...
+        def HasField(self, field_name: typing.Literal["adamw", b"adamw", "type", b"type"]) -> builtins.bool: ...
+        def ClearField(self, field_name: typing.Literal["adamw", b"adamw", "type", b"type"]) -> None: ...
+        def WhichOneof(self, oneof_group: typing.Literal["type", b"type"]) -> typing.Literal["adamw"] | None: ...
+
+    SEED_FIELD_NUMBER: builtins.int
+    LR_FIELD_NUMBER: builtins.int
+    N_SAMPLES_FIELD_NUMBER: builtins.int
+    ADDITIONAL_ARGUMENTS_FIELD_NUMBER: builtins.int
+    LORA_FIELD_NUMBER: builtins.int
+    PROMPT_TUNE_FIELD_NUMBER: builtins.int
+    SCHEDULER_FIELD_NUMBER: builtins.int
+    OPTIMIZER_FIELD_NUMBER: builtins.int
+    seed: builtins.int
+    lr: builtins.float
+    n_samples: builtins.int
+    additional_arguments: builtins.str
+    @property
+    def lora(self) -> yandex.cloud.ai.tuning.v1.tuning_types_pb2.TuningTypeLora: ...
+    @property
+    def prompt_tune(self) -> yandex.cloud.ai.tuning.v1.tuning_types_pb2.TuningTypePromptTune: ...
+    @property
+    def scheduler(self) -> global___TextClassificationMulticlassParams.Scheduler: ...
+    @property
+    def optimizer(self) -> global___TextClassificationMulticlassParams.Optimizer: ...
+    def __init__(
+        self,
+        *,
+        seed: builtins.int = ...,
+        lr: builtins.float = ...,
+        n_samples: builtins.int = ...,
+        additional_arguments: builtins.str = ...,
+        lora: yandex.cloud.ai.tuning.v1.tuning_types_pb2.TuningTypeLora | None = ...,
+        prompt_tune: yandex.cloud.ai.tuning.v1.tuning_types_pb2.TuningTypePromptTune | None = ...,
+        scheduler: global___TextClassificationMulticlassParams.Scheduler | None = ...,
+        optimizer: global___TextClassificationMulticlassParams.Optimizer | None = ...,
+    ) -> None: ...
+    def HasField(self, field_name: typing.Literal["lora", b"lora", "optimizer", b"optimizer", "prompt_tune", b"prompt_tune", "scheduler", b"scheduler", "tuning_type", b"tuning_type"]) -> builtins.bool: ...
+    def ClearField(self, field_name: typing.Literal["additional_arguments", b"additional_arguments", "lora", b"lora", "lr", b"lr", "n_samples", b"n_samples", "optimizer", b"optimizer", "prompt_tune", b"prompt_tune", "scheduler", b"scheduler", "seed", b"seed", "tuning_type", b"tuning_type"]) -> None: ...
+    def WhichOneof(self, oneof_group: typing.Literal["tuning_type", b"tuning_type"]) -> typing.Literal["lora", "prompt_tune"] | None: ...
+
+global___TextClassificationMulticlassParams = TextClassificationMulticlassParams
+
+@typing.final
+class GetMetricsUrlRequest(google.protobuf.message.Message):
+    DESCRIPTOR: google.protobuf.descriptor.Descriptor
+
+    TASK_ID_FIELD_NUMBER: builtins.int
+    task_id: builtins.str
+    def __init__(
+        self,
+        *,
+        task_id: builtins.str = ...,
+    ) -> None: ...
+    def ClearField(self, field_name: typing.Literal["task_id", b"task_id"]) -> None: ...
+
+global___GetMetricsUrlRequest = GetMetricsUrlRequest
+
+@typing.final
+class GetMetricsUrlResponse(google.protobuf.message.Message):
+    DESCRIPTOR: google.protobuf.descriptor.Descriptor
+
+    LOAD_URL_FIELD_NUMBER: builtins.int
+    load_url: builtins.str
+    def __init__(
+        self,
+        *,
+        load_url: builtins.str = ...,
+    ) -> None: ...
+    def ClearField(self, field_name: typing.Literal["load_url", b"load_url"]) -> None: ...
+
+global___GetMetricsUrlResponse = GetMetricsUrlResponse
+
+@typing.final
+class GetOptionsRequest(google.protobuf.message.Message):
+    DESCRIPTOR: google.protobuf.descriptor.Descriptor
+
+    TASK_ID_FIELD_NUMBER: builtins.int
+    task_id: builtins.str
+    def __init__(
+        self,
+        *,
+        task_id: builtins.str = ...,
+    ) -> None: ...
+    def ClearField(self, field_name: typing.Literal["task_id", b"task_id"]) -> None: ...
+
+global___GetOptionsRequest = GetOptionsRequest
+
+@typing.final
+class GetOptionsResponse(google.protobuf.message.Message):
+    DESCRIPTOR: google.protobuf.descriptor.Descriptor
+
+    TASK_ID_FIELD_NUMBER: builtins.int
+    BASE_MODEL_URI_FIELD_NUMBER: builtins.int
+    TRAIN_DATASETS_FIELD_NUMBER: builtins.int
+    VALIDATION_DATASETS_FIELD_NUMBER: builtins.int
+    TEXT_TO_TEXT_COMPLETION_FIELD_NUMBER: builtins.int
+    TEXT_CLASSIFICATION_MULTILABEL_FIELD_NUMBER: builtins.int
+    TEXT_CLASSIFICATION_MULTICLASS_FIELD_NUMBER: builtins.int
+    task_id: builtins.str
+    base_model_uri: builtins.str
+    @property
+    def train_datasets(self) -> google.protobuf.internal.containers.RepeatedCompositeFieldContainer[global___TuningRequest.WeightedDataset]: ...
+    @property
+    def validation_datasets(self) -> google.protobuf.internal.containers.RepeatedCompositeFieldContainer[global___TuningRequest.WeightedDataset]: ...
+    @property
+    def text_to_text_completion(self) -> global___TextToTextCompletionTuningParams: ...
+    @property
+    def text_classification_multilabel(self) -> global___TextClassificationMultilabelParams: ...
+    @property
+    def text_classification_multiclass(self) -> global___TextClassificationMulticlassParams: ...
+    def __init__(
+        self,
+        *,
+        task_id: builtins.str = ...,
+        base_model_uri: builtins.str = ...,
+        train_datasets: collections.abc.Iterable[global___TuningRequest.WeightedDataset] | None = ...,
+        validation_datasets: collections.abc.Iterable[global___TuningRequest.WeightedDataset] | None = ...,
+        text_to_text_completion: global___TextToTextCompletionTuningParams | None = ...,
+        text_classification_multilabel: global___TextClassificationMultilabelParams | None = ...,
+        text_classification_multiclass: global___TextClassificationMulticlassParams | None = ...,
+    ) -> None: ...
+    def HasField(self, field_name: typing.Literal["text_classification_multiclass", b"text_classification_multiclass", "text_classification_multilabel", b"text_classification_multilabel", "text_to_text_completion", b"text_to_text_completion", "tuning_params", b"tuning_params"]) -> builtins.bool: ...
+    def ClearField(self, field_name: typing.Literal["base_model_uri", b"base_model_uri", "task_id", b"task_id", "text_classification_multiclass", b"text_classification_multiclass", "text_classification_multilabel", b"text_classification_multilabel", "text_to_text_completion", b"text_to_text_completion", "train_datasets", b"train_datasets", "tuning_params", b"tuning_params", "validation_datasets", b"validation_datasets"]) -> None: ...
+    def WhichOneof(self, oneof_group: typing.Literal["tuning_params", b"tuning_params"]) -> typing.Literal["text_to_text_completion", "text_classification_multilabel", "text_classification_multiclass"] | None: ...
+
+global___GetOptionsResponse = GetOptionsResponse