Source code for easydel.trainers.trainer.modeling_output
# Copyright 2025 The EasyDeL Author @erfanzar (Erfan Zare Chavoshi).
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# https://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""Training output data structures.
This module defines the output types returned by training operations,
including the final state and associated metadata from training runs.
"""
from __future__ import annotations
import typing as tp
import jax
from eformer.pytree import auto_pytree
if tp.TYPE_CHECKING:
from easydel.infra.base_state import EasyDeLState
else:
EasyDeLState = tp.Any
CallFN: tp.TypeAlias = tp.Any | tp.Mapping[str, tp.Callable] | dict[str, tp.Callable]
[docs]@auto_pytree
class TrainerOutput:
"""Output from a training run.
Contains the final model state and metadata from the training process,
including checkpointing information and utility functions for state manipulation.
Attributes:
state: The final EasyDeLState after training completion.
mesh: The JAX sharding mesh used during training, if any.
checkpoint_manager: Manager object for handling model checkpoints.
gather_fns: Functions for gathering sharded parameters to host.
shard_fns: Functions for sharding parameters across devices.
last_save_file_name: Name of the most recently saved checkpoint file.
checkpoint_path: Full path to the checkpoint directory.
"""
state: EasyDeLState
mesh: jax.sharding.Mesh | None
checkpoint_manager: tp.Any
gather_fns: CallFN | None = None
shard_fns: CallFN | None = None
last_save_file_name: str | None = None
checkpoint_path: str | None = None