Spaces:
Running
Running
File size: 3,062 Bytes
e81015c |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 |
# Copyright 2025 the LlamaFactory team.
#
# 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
#
# http://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.
from types import MethodType
from typing import TYPE_CHECKING, Optional
import torch
from transformers import Trainer
from typing_extensions import override
from ...extras.packages import is_transformers_version_greater_than
from ..callbacks import SaveProcessorCallback
from ..trainer_utils import create_custom_optimizer, create_custom_scheduler
if TYPE_CHECKING:
from transformers import ProcessorMixin
from ...hparams import FinetuningArguments
class CustomTrainer(Trainer):
r"""Inherit Trainer for custom optimizer."""
def __init__(
self, finetuning_args: "FinetuningArguments", processor: Optional["ProcessorMixin"], **kwargs
) -> None:
if is_transformers_version_greater_than("4.46"):
kwargs["processing_class"] = kwargs.pop("tokenizer")
super().__init__(**kwargs)
if processor is not None:
# avoid wrong loss under gradient accumulation
# https://github.com/huggingface/transformers/pull/36044#issuecomment-2746657112
self.model_accepts_loss_kwargs = False
self.finetuning_args = finetuning_args
if processor is not None:
self.add_callback(SaveProcessorCallback(processor))
if finetuning_args.use_badam:
from badam import BAdamCallback, clip_grad_norm_old_version # type: ignore
self.accelerator.clip_grad_norm_ = MethodType(clip_grad_norm_old_version, self.accelerator)
self.add_callback(BAdamCallback)
@override
def create_optimizer(self) -> "torch.optim.Optimizer":
if self.optimizer is None:
self.optimizer = create_custom_optimizer(self.model, self.args, self.finetuning_args)
return super().create_optimizer()
@override
def create_scheduler(
self, num_training_steps: int, optimizer: Optional["torch.optim.Optimizer"] = None
) -> "torch.optim.lr_scheduler.LRScheduler":
create_custom_scheduler(self.args, num_training_steps, optimizer)
return super().create_scheduler(num_training_steps, optimizer)
@override
def _get_train_sampler(self) -> Optional["torch.utils.data.Sampler"]:
if self.finetuning_args.disable_shuffling:
return torch.utils.data.SequentialSampler(self.train_dataset)
return super()._get_train_sampler()
@override
def compute_loss(self, model, inputs, *args, **kwargs):
return super().compute_loss(model, inputs, *args, **kwargs)
|