DPO Trainer

TRL supports the DPO Trainer for training language models from preference data, as described in the paper Direct Preference Optimization: Your Language Model is Secretly a Reward Model by Rafailov et al., 2023. For a full example have a look at examples/dpo.py.

The first step as always is to train your SFT model, to ensure the data we train on is in-distribution for the DPO algorithm.

Expected dataset format

The DPO trainer expects a very specific format for the dataset. Since the model will be trained to directly optimize the preference of which sentence is the most relevant, given two sentences. We provide an example from the Anthropic/hh-rlhf dataset below:

Therefore the final dataset object should contain these 3 entries if you use the default DPODataCollatorWithPadding data collator. The entries should be named:

prompt
chosen
rejected

for example:

dpo_dataset_dict = {
    "prompt": [
        "hello",
        "how are you",
        "What is your name?",
        "What is your name?",
        "Which is the best programming language?",
        "Which is the best programming language?",
        "Which is the best programming language?",
    ],
    "chosen": [
        "hi nice to meet you",
        "I am fine",
        "My name is Mary",
        "My name is Mary",
        "Python",
        "Python",
        "Java",
    ],
    "rejected": [
        "leave me alone",
        "I am not fine",
        "Whats it to you?",
        "I dont have a name",
        "Javascript",
        "C++",
        "C++",
    ],
}

where the prompt contains the context inputs, chosen contains the corresponding chosen responses and rejected contains the corresponding negative (rejected) responses. As can be seen a prompt can have multiple responses and this is reflected in the entries being repeated in the dictionary’s value arrays.

Expected model format

The DPO trainer expects a model of AutoModelForCausalLM, compared to PPO that expects AutoModelForCausalLMWithValueHead for the value function.

Using the DPOTrainer

For a detailed example have a look at the examples/dpo.py script. At a high level we need to initialize the DPOTrainer with a model we wish to train, a reference ref_model which we will use to calculate the implicit rewards of the preferred and rejected response, the beta refers to the hyperparameter of the implicit reward, and the dataset contains the 3 entries listed above. Note that the model and ref_model need to have the same architecture (ie decoder only or encoder-decoder).

 dpo_trainer = DPOTrainer(
    model,
    model_ref,
    args=training_args,
    beta=0.1,
    train_dataset=train_dataset,
    tokenizer=tokenizer,
)

After this one can then call:

dpo_trainer.train()

Note that the beta is the temperature parameter for the DPO loss, typically something in the range of 0.1 to 0.5. We ignore the reference model as beta -> 0.

Loss function

Given the preference data, we can fit a binary classifier according to the Bradley-Terry model and in fact the DPO authors propose the sigmoid loss on the normalized likelihood via the logsigmoid to fit a logistic regression. The RSO authors propose to use a hinge loss on the normalized likelihood from the SLiC paper. The DPOTrainer can be switched to this loss via the loss_type="hinge" argument and the beta in this case is the reciprocal of the margin.

Logging

While training and evaluating we record the following reward metrics:

rewards/chosen: the mean difference between the log probabilities of the policy model and the reference model for the chosen responses scaled by beta
rewards/rejected: the mean difference between the log probabilities of the policy model and the reference model for the rejected responses scaled by beta
rewards/accuracies: mean of how often the chosen rewards are > than the corresponding rejected rewards
rewards/margins: the mean difference between the chosen and corresponding rejected rewards

DPOTrainer

class trl.DPOTrainer

< source >

( model: typing.Union[transformers.modeling_utils.PreTrainedModel, torch.nn.modules.module.Module, str] = None ref_model: typing.Union[transformers.modeling_utils.PreTrainedModel, torch.nn.modules.module.Module, str, NoneType] = None beta: float = 0.1 loss_type: typing.Literal['sigmoid', 'hinge'] = 'sigmoid' args: TrainingArguments = None data_collator: typing.Optional[DataCollator] = None label_pad_token_id: int = -100 padding_value: int = 0 truncation_mode: str = 'keep_end' train_dataset: typing.Optional[datasets.arrow_dataset.Dataset] = None eval_dataset: typing.Union[datasets.arrow_dataset.Dataset, typing.Dict[str, datasets.arrow_dataset.Dataset], NoneType] = None tokenizer: typing.Optional[transformers.tokenization_utils_base.PreTrainedTokenizerBase] = None model_init: typing.Union[typing.Callable[[], transformers.modeling_utils.PreTrainedModel], NoneType] = None callbacks: typing.Optional[typing.List[transformers.trainer_callback.TrainerCallback]] = None optimizers: typing.Tuple[torch.optim.optimizer.Optimizer, torch.optim.lr_scheduler.LambdaLR] = (None, None) preprocess_logits_for_metrics: typing.Union[typing.Callable[[torch.Tensor, torch.Tensor], torch.Tensor], NoneType] = None max_length: typing.Optional[int] = None max_prompt_length: typing.Optional[int] = None max_target_length: typing.Optional[int] = None peft_config: typing.Optional[typing.Dict] = None is_encoder_decoder: typing.Optional[bool] = None disable_dropout: bool = True generate_during_eval: bool = False compute_metrics: typing.Union[typing.Callable[[transformers.trainer_utils.EvalLoopOutput], typing.Dict], NoneType] = None model_init_kwargs: typing.Optional[typing.Dict] = None ref_model_init_kwargs: typing.Optional[typing.Dict] = None )

Parameters

model (transformers.PreTrainedModel) — The model to train, preferably an AutoModelForSequenceClassification.
ref_model (PreTrainedModelWrapper) — Hugging Face transformer model with a casual language modelling head. Used for implicit reward computation and loss. If no reference model is provided, the trainer will create a reference model with the same architecture as the model to be optimized.
beta (float, defaults to 0.1) — The beta factor in DPO loss. Higher beta means less divergence from the initial policy.
loss_type (str, defaults to "sigmoid") — The type of DPO loss to use. Either "sigmoid" the default DPO loss or "hinge" loss from SLiC paper.
args (transformers.TrainingArguments) — The arguments to use for training.
data_collator (transformers.DataCollator) — The data collator to use for training. If None is specified, the default data collator (DPODataCollatorWithPadding) will be used which will pad the sequences to the maximum length of the sequences in the batch, given a dataset of paired sequences.
label_pad_token_id (int, defaults to -100) — The label pad token id. This argument is required if you want to use the default data collator.
padding_value (int, defaults to 0) — The padding value. This argument is required if you want to use the default data collator.
truncation_mode (str, defaults to keep_end) — The truncation mode to use, either keep_end or keep_start. This argument is required if you want to use the default data collator.
train_dataset (datasets.Dataset) — The dataset to use for training.
eval_dataset (datasets.Dataset) — The dataset to use for evaluation.
tokenizer (transformers.PreTrainedTokenizerBase) — The tokenizer to use for training. This argument is required if you want to use the default data collator.
model_init (Callable[[], transformers.PreTrainedModel]) — The model initializer to use for training. If None is specified, the default model initializer will be used.
callbacks (List[transformers.TrainerCallback]) — The callbacks to use for training.
optimizers (Tuple[torch.optim.Optimizer, torch.optim.lr_scheduler.LambdaLR]) — The optimizer and scheduler to use for training.
preprocess_logits_for_metrics (Callable[[torch.Tensor, torch.Tensor], torch.Tensor]) — The function to use to preprocess the logits before computing the metrics.
max_length (int, defaults to None) — The maximum length of the sequences in the batch. This argument is required if you want to use the default data collator.
max_prompt_length (int, defaults to None) — The maximum length of the prompt. This argument is required if you want to use the default data collator.
max_target_length (int, defaults to None) — The maximum length of the target. This argument is required if you want to use the default data collator and your model is an encoder-decoder.
peft_config (Dict, defaults to None) — The PEFT configuration to use for training. If you pass a PEFT configuration, the model will be wrapped in a PEFT model.
is_encoder_decoder (Optional[bool], optional, defaults to None) — If no model is provided, we need to know if the model_init returns an encoder-decoder.
disable_dropout (bool, defaults to True) — Whether or not to disable dropouts in model and ref_model.
generate_during_eval (bool, defaults to False) — Whether to sample and log generations during evaluation step.
compute_metrics (Callable[[EvalPrediction], Dict], optional) — The function to use to compute the metrics. Must take a EvalPrediction and return a dictionary string to metric values. model_init_kwargs — (Optional[Dict], optional): Dict of Optional kwargs to pass when instantiating the model from a string ref_model_init_kwargs — (Optional[Dict], optional): Dict of Optional kwargs to pass when instantiating the ref model from a string

Initialize DPOTrainer.

concatenated_forward

< source >

( model: Module batch: typing.Dict[str, typing.Union[typing.List, torch.LongTensor]] )

Run the given model on the given batch of inputs, concatenating the chosen and rejected inputs together.

We do this to avoid doing two forward passes, because it’s faster for FSDP.

concatenated_inputs

< source >

( batch: typing.Dict[str, typing.Union[typing.List, torch.LongTensor]] )

Concatenate the chosen and rejected inputs into a single tensor.

dpo_loss

< source >

( policy_chosen_logps: FloatTensor policy_rejected_logps: FloatTensor reference_chosen_logps: FloatTensor reference_rejected_logps: FloatTensor reference_free: bool = False ) → A tuple of three tensors

Returns

A tuple of three tensors

(losses, chosen_rewards, rejected_rewards). The losses tensor contains the DPO loss for each example in the batch. The chosen_rewards and rejected_rewards tensors contain the rewards for the chosen and rejected responses, respectively.

Compute the DPO loss for a batch of policy and reference model log probabilities.

evaluation_loop

< source >

( dataloader: DataLoader description: str prediction_loss_only: typing.Optional[bool] = None ignore_keys: typing.Optional[typing.List[str]] = None metric_key_prefix: str = 'eval' )

Overriding built-in evaluation loop to store metrics for each batch. Prediction/evaluation loop, shared by Trainer.evaluate() and Trainer.predict().

Works both with or without labels.

get_batch_metrics

< source >

( model batch: typing.Dict[str, typing.Union[typing.List, torch.LongTensor]] train_eval: typing.Literal['train', 'eval'] = 'train' )

Compute the DPO loss and other metrics for the given batch of inputs for train or test.

get_batch_samples

< source >

( model batch: typing.Dict[str, torch.LongTensor] )

Generate samples from the model and reference model for the given batch of inputs.

log

< source >

( logs: typing.Dict[str, float] )

Parameters

logs (Dict[str, float]) — The values to log.

Log logs on the various objects watching training, including stored metrics.