trainer.py

# Copyright 2024 HuggingFace Inc. and the LlamaFactory team.
#
# This code is inspired by the HuggingFace's transformers library.
# https://github.com/huggingface/transformers/blob/v4.40.0/src/transformers/trainer.py
#
# 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.

import json
import os
from types import MethodType
from typing import TYPE_CHECKING, Dict, List, Optional, Tuple, Union

import torch
from transformers import Trainer
from typing_extensions import override

from ...extras import logging
from ...extras.packages import is_transformers_version_equal_to_4_46, is_transformers_version_greater_than
from ..callbacks import FixValueHeadModelCallback, SaveProcessorCallback
from ..trainer_utils import create_custom_optimizer, create_custom_scheduler


if TYPE_CHECKING:
    from transformers import PreTrainedModel, ProcessorMixin
    from transformers.trainer import PredictionOutput

    from ...hparams import FinetuningArguments


logger = logging.get_logger(__name__)


class PairwiseTrainer(Trainer):
    r"""
    Inherits Trainer to compute pairwise loss.
    """

    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)
        self.model_accepts_loss_kwargs = False  # overwrite trainer's default behavior
        self.finetuning_args = finetuning_args
        self.can_return_loss = True  # override property to return eval_loss
        self.add_callback(FixValueHeadModelCallback)

        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: "PreTrainedModel", inputs: Dict[str, "torch.Tensor"], return_outputs: bool = False, **kwargs
    ) -> Union["torch.Tensor", Tuple["torch.Tensor", List["torch.Tensor"]]]:
        r"""
        Computes pairwise loss. The first n examples are chosen and the last n examples are rejected.

        Subclass and override to inject custom behavior.

        Note that the first element will be removed from the output tuple.
        See: https://github.com/huggingface/transformers/blob/v4.40.0/src/transformers/trainer.py#L3842
        """
        _, _, values = model(**inputs, output_hidden_states=True, return_dict=True, use_cache=False)
        batch_size = inputs["input_ids"].size(0) // 2
        chosen_masks, rejected_masks = torch.split(inputs["attention_mask"], batch_size, dim=0)
        chosen_rewards, rejected_rewards = torch.split(values, batch_size, dim=0)
        chosen_scores = chosen_rewards.gather(dim=-1, index=(chosen_masks.sum(dim=-1, keepdim=True) - 1))
        rejected_scores = rejected_rewards.gather(dim=-1, index=(rejected_masks.sum(dim=-1, keepdim=True) - 1))
        chosen_scores, rejected_scores = chosen_scores.squeeze(), rejected_scores.squeeze()

        loss = -torch.nn.functional.logsigmoid(chosen_scores.float() - rejected_scores.float()).mean()

        if is_transformers_version_equal_to_4_46() and kwargs.get("num_items_in_batch"):
            loss /= self.args.gradient_accumulation_steps  # fixes the loss value for transformers 4.46.0-4.46.1

        if return_outputs:
            return loss, (loss, chosen_scores, rejected_scores)
        else:
            return loss

    def save_predictions(self, predict_results: "PredictionOutput") -> None:
        r"""
        Saves model predictions to `output_dir`.

        A custom behavior that not contained in Seq2SeqTrainer.
        """
        if not self.is_world_process_zero():
            return

        output_prediction_file = os.path.join(self.args.output_dir, "generated_predictions.jsonl")
        logger.info_rank0(f"Saving prediction results to {output_prediction_file}")
        chosen_scores, rejected_scores = predict_results.predictions

        with open(output_prediction_file, "w", encoding="utf-8") as writer:
            res: List[str] = []
            for c_score, r_score in zip(chosen_scores, rejected_scores):
                res.append(json.dumps({"chosen": round(float(c_score), 2), "rejected": round(float(r_score), 2)}))

            writer.write("\n".join(res))