Data Utilities

prepare_multimodal_messages

trl.prepare_multimodal_messages

( messages: list images: list ) → list[dict[str, Any]]

Parameters

messages (list[dict[str, Any]]) — Messages with "role" and "content". Content may be a raw string before transformation. List of messages a "role" key ("system", "user", or "assistant") and a "content" key containing either a string or a list of structured blocks if already prepared.
images (list) — List of image objects to insert.

Returns

list[dict[str, Any]]

A deep-copied list of messages where every "content" value is a list of structured content blocks, and all "image" placeholders are populated with the corresponding image objects.

Convert messages into a structured multimodal format and inject the provided images into the message contents.

Notes:

When the input messages isn’t already in the structured format, (i.e., all "content" values are strings), the function transforms them into the structured format by wrapping text in {"type": "text", "text": ...} and inserting {"type": "image"} placeholders for the images before the first user message.
When the input messages is already in the structured format (i.e., all "content" values are lists of structured blocks), the function only fills in the actual images in the existing {"type": "image"} placeholders. If the number of placeholders does not match the number of provided images, an error is raised.

Example:

# Input
[
    {"role": "user", "content": "What's in this image?"},
    {"role": "assistant", "content": "It looks like a cat."},
]

# Output, one image provided
[
    {"role": "user", "content": [{"type": "image", "image": <PIL.Image.Image>}, {"type": "text", "text": "What's in this image?"}]},
    {"role": "assistant", "content": [{"type": "text", "text": "It looks like a cat."}]},
]

prepare_multimodal_messages_vllm

trl.prepare_multimodal_messages_vllm

< source >

( messages: list ) → list[dict[str, Any]]

Parameters

messages (list[dict[str, Any]]) — Messages with "role" and "content". Content is expected to be a list of structured blocks.

Returns

list[dict[str, Any]]

A deep-copied list of messages compatible with vLLM’s expected input format.

Convert structured multimodal messages into a format compatible with vLLM. Replaces "type": "image" blocks with "type": "image_pil" blocks, and "image": Image with "image_pil": Image.

Example:

# Input
[{"role": "user", "content": [{"type": "image", "image": <PIL.Image.Image>}, {"type": "text", "text": "What's in this image?"}]}]

# Output
[{"role": "user", "content": [{"type": "image_pil", "image_pil": <PIL.Image.Image>}, {"type": "text", "text": "What's in this image?"}]}]

is_conversational

trl.is_conversational

< source >

( example: dict ) → bool

Parameters

example (dict[str, Any]) — A single data entry of a dataset. The example can have different keys depending on the dataset type.

Returns

bool

True if the data is in a conversational format, False otherwise.

Check if the example is in a conversational format.

Examples:

>>> example = {"prompt": [{"role": "user", "content": "What color is the sky?"}]}
>>> is_conversational(example)
True

>>> example = {"prompt": "The sky is"}
>>> is_conversational(example)
False

is_conversational_from_value

trl.is_conversational_from_value

< source >

( example: dict ) → bool

Parameters

example (dict[str, Any]) — A single data entry of a dataset. The example can have different keys depending on the dataset type.

Returns

bool

True if the data is in a conversational Chatformat, False otherwise.

Check if the example is in a conversational format (from/value). Note that this format isn’t recommended. Prefer the ChatML format (role/content)

Examples:

>>> example = {"conversations": [{"from": "user", "value": "What color is the sky?"}]}
>>> is_conversational_from_value(example)
True

>>> example = {"conversations": [{"role": "user", "content": "What color is the sky?"}]}
>>> is_conversational_from_value(example)
False

>>> example = {"conversations": "The sky is"}
>>> is_conversational_from_value(example)
False

apply_chat_template

trl.apply_chat_template

< source >

( example: dict tokenizer: typing.Union[transformers.tokenization_utils_base.PreTrainedTokenizerBase, transformers.processing_utils.ProcessorMixin] tools: typing.Optional[list[typing.Union[dict, typing.Callable]]] = None **template_kwargs )

Apply a chat template to a conversational example along with the schema for a list of functions in tools.

For more details, see maybe_apply_chat_template().

maybe_apply_chat_template

trl.maybe_apply_chat_template

< source >

( example: dict tokenizer: PreTrainedTokenizerBase tools: typing.Optional[list[typing.Union[dict, typing.Callable]]] = None **template_kwargs: typing.Any ) → dict[str, str]

Parameters

example (dict[str, list[dict[str, str]]) — Dictionary representing a single data entry of a conversational dataset. Each data entry can have different keys depending on the dataset type. The supported dataset types are:
- Language modeling dataset: "messages".
- Prompt-only dataset: "prompt".
- Prompt-completion dataset: "prompt" and "completion".
- Preference dataset: "prompt", "chosen", and "rejected".
- Preference dataset with implicit prompt: "chosen" and "rejected".
- Unpaired preference dataset: "prompt", "completion", and "label".
For keys "messages", "prompt", "chosen", "rejected", and "completion", the values are lists of messages, where each message is a dictionary with keys "role" and "content". Additionally, the example may contain a "chat_template_kwargs" key, which is a dictionary of additional keyword arguments to pass to the chat template renderer.
tokenizer (PreTrainedTokenizerBase) — Tokenizer to apply the chat template with.
tools (list[Union[dict, Callable]], optional) — A list of tools (callable functions) that will be accessible to the model. If the template does not support function calling, this argument will have no effect.
**template_kwargs (Any, optional) — Additional kwargs to pass to the template renderer. Will be accessible by the chat template.

Returns

dict[str, str]

Formatted example with the chat template applied.

If the example is in a conversational format, apply a chat template to it.

Notes:

This function does not alter the keys, except for Language modeling dataset, where "messages" is replaced by "text".
In case of prompt-only data, if the last role is "user", the generation prompt is added to the prompt. Else, if the last role is "assistant", the final message is continued.

Example:

>>> from transformers import AutoTokenizer

>>> tokenizer = AutoTokenizer.from_pretrained("microsoft/Phi-3-mini-128k-instruct")
>>> example = {
...     "prompt": [{"role": "user", "content": "What color is the sky?"}],
...     "completion": [{"role": "assistant", "content": "It is blue."}],
... }
>>> apply_chat_template(example, tokenizer)
{'prompt': '<|user|>\nWhat color is the sky?<|end|>\n<|assistant|>\n', 'completion': 'It is blue.<|end|>\n'}

maybe_convert_to_chatml

trl.maybe_convert_to_chatml

< source >

( example: dict ) → dict[str, list]

Parameters

example (dict[str, list]) — A single data entry containing a list of messages.

Returns

dict[str, list]

Example reformatted to ChatML style.

Convert a conversational dataset with fields from and value to ChatML format.

This function modifies conversational data to align with OpenAI’s ChatML format:

Replaces the key "from" with "role" in message dictionaries.
Replaces the key "value" with "content" in message dictionaries.
Renames "conversations" to "messages" for consistency with ChatML.

Example:

>>> from trl import maybe_convert_to_chatml

>>> example = {
...     "conversations": [
...         {"from": "user", "value": "What color is the sky?"},
...         {"from": "assistant", "value": "It is blue."},
...     ]
... }
>>> maybe_convert_to_chatml(example)
{'messages': [{'role': 'user', 'content': 'What color is the sky?'},
              {'role': 'assistant', 'content': 'It is blue.'}]}

extract_prompt

trl.extract_prompt

< source >

( example: dict )

Extracts the shared prompt from a preference data example, where the prompt is implicit within both the chosen and rejected completions.

For more details, see maybe_extract_prompt().

maybe_extract_prompt

trl.maybe_extract_prompt

< source >

( example: dict ) → dict[str, list]

Parameters

example (dict[str, list]) — A dictionary representing a single data entry in the preference dataset. It must contain the keys "chosen" and "rejected", where each value is either conversational or standard (str).

Returns

dict[str, list]

A dictionary containing:

"prompt": The longest common prefix between the “chosen” and “rejected” completions.
"chosen": The remainder of the “chosen” completion, with the prompt removed.
"rejected": The remainder of the “rejected” completion, with the prompt removed.

Extracts the shared prompt from a preference data example, where the prompt is implicit within both the chosen and rejected completions.

If the example already contains a "prompt" key, the function returns the example as is. Else, the function identifies the longest common sequence (prefix) of conversation turns between the “chosen” and “rejected” completions and extracts this as the prompt. It then removes this prompt from the respective “chosen” and “rejected” completions.

Examples:

>>> example = {
...     "chosen": [
...         {"role": "user", "content": "What color is the sky?"},
...         {"role": "assistant", "content": "It is blue."},
...     ],
...     "rejected": [
...         {"role": "user", "content": "What color is the sky?"},
...         {"role": "assistant", "content": "It is green."},
...     ],
... }
>>> extract_prompt(example)
{'prompt': [{'role': 'user', 'content': 'What color is the sky?'}],
 'chosen': [{'role': 'assistant', 'content': 'It is blue.'}],
 'rejected': [{'role': 'assistant', 'content': 'It is green.'}]}

Or, with the map method of Dataset:

>>> from trl import extract_prompt
>>> from datasets import Dataset

>>> dataset_dict = {
...     "chosen": [
...         [
...             {"role": "user", "content": "What color is the sky?"},
...             {"role": "assistant", "content": "It is blue."},
...         ],
...         [
...             {"role": "user", "content": "Where is the sun?"},
...             {"role": "assistant", "content": "In the sky."},
...         ],
...     ],
...     "rejected": [
...         [
...             {"role": "user", "content": "What color is the sky?"},
...             {"role": "assistant", "content": "It is green."},
...         ],
...         [
...             {"role": "user", "content": "Where is the sun?"},
...             {"role": "assistant", "content": "In the sea."},
...         ],
...     ],
... }
>>> dataset = Dataset.from_dict(dataset_dict)
>>> dataset = dataset.map(extract_prompt)
>>> dataset[0]
{'prompt': [{'role': 'user', 'content': 'What color is the sky?'}],
 'chosen': [{'role': 'assistant', 'content': 'It is blue.'}],
 'rejected': [{'role': 'assistant', 'content': 'It is green.'}]}

unpair_preference_dataset

trl.unpair_preference_dataset

< source >

( dataset: ~DatasetType num_proc: typing.Optional[int] = None desc: typing.Optional[str] = None ) → Dataset

Parameters

dataset (Dataset or DatasetDict) — Preference dataset to unpair. The dataset must have columns "chosen", "rejected" and optionally "prompt".
num_proc (int, optional) — Number of processes to use for processing the dataset.
desc (str, optional) — Meaningful description to be displayed alongside with the progress bar while mapping examples.

Returns

Dataset

The unpaired preference dataset.

Unpair a preference dataset.

Example:

>>> from datasets import Dataset

>>> dataset_dict = {
...     "prompt": ["The sky is", "The sun is"],
...     "chosen": [" blue.", "in the sky."],
...     "rejected": [" green.", " in the sea."],
... }
>>> dataset = Dataset.from_dict(dataset_dict)
>>> dataset = unpair_preference_dataset(dataset)
>>> dataset
Dataset({
    features: ['prompt', 'completion', 'label'],
    num_rows: 4
})

>>> dataset[0]
{'prompt': 'The sky is', 'completion': ' blue.', 'label': True}

maybe_unpair_preference_dataset

trl.maybe_unpair_preference_dataset

< source >

( dataset: ~DatasetType num_proc: typing.Optional[int] = None desc: typing.Optional[str] = None ) → Dataset or DatasetDict

Parameters

dataset (Dataset or DatasetDict) — Preference dataset to unpair. The dataset must have columns "chosen", "rejected" and optionally "prompt".
num_proc (int, optional) — Number of processes to use for processing the dataset.
desc (str, optional) — Meaningful description to be displayed alongside with the progress bar while mapping examples.

Returns

Dataset or DatasetDict

The unpaired preference dataset if it was paired, otherwise the original dataset.

Unpair a preference dataset if it is paired.

Example:

>>> from datasets import Dataset

>>> dataset_dict = {
...     "prompt": ["The sky is", "The sun is"],
...     "chosen": [" blue.", "in the sky."],
...     "rejected": [" green.", " in the sea."],
... }
>>> dataset = Dataset.from_dict(dataset_dict)
>>> dataset = unpair_preference_dataset(dataset)
>>> dataset
Dataset({
    features: ['prompt', 'completion', 'label'],
    num_rows: 4
})

>>> dataset[0]
{'prompt': 'The sky is', 'completion': ' blue.', 'label': True}

pack_dataset

trl.pack_dataset

< source >

( dataset: ~DatasetType seq_length: int strategy: str = 'bfd' map_kwargs: typing.Optional[dict[str, typing.Any]] = None ) → Dataset or DatasetDict

Parameters

dataset (Dataset or DatasetDict) — Dataset to pack
seq_length (int) — Target sequence length to pack to.
strategy (str, optional, defaults to "bfd") — Packing strategy to use. Can be either:
- "bfd" (Best Fit Decreasing): Slower but preserves sequence boundaries. Sequences are never cut in the middle.
- "wrapped": Faster but more aggressive. Ignores sequence boundaries and will cut sequences in the middle to completely fill each packed sequence with data.
map_kwargs (dict, optional) — Additional keyword arguments to pass to the dataset’s map method when packing examples.

Returns

Dataset or DatasetDict

The dataset with packed sequences. The number of examples may decrease as sequences are combined.

Pack sequences in a dataset into chunks of size seq_length.

Example:

>>> from datasets import Dataset
>>> from trl import pack_dataset

>>> examples = {
...     "input_ids": [[1, 2, 3], [4, 5], [6, 7, 8], [9]],
...     "attention_mask": [[1, 1, 0], [1, 0], [1, 0, 0], [1]],
... }
>>> dataset = Dataset.from_dict(examples)
>>> packed_dataset = pack_dataset(dataset, seq_length=4, strategy="bfd")
>>> packed_dataset[:]
{'input_ids': [[1, 2, 3, 9], [6, 7, 8], [4, 5]],
'attention_mask': [[1, 1, 0, 1], [1, 0, 0], [1, 0]],
'seq_lengths': [[3, 1], [3], [2]]}

truncate_dataset

trl.truncate_dataset

< source >

( dataset: ~DatasetType max_length: int map_kwargs: typing.Optional[dict[str, typing.Any]] = None ) → Dataset or DatasetDict

Parameters

dataset (Dataset or DatasetDict) — Dataset to truncate.
max_length (int) — Maximum sequence length to truncate to.
map_kwargs (dict, optional) — Additional keyword arguments to pass to the dataset’s map method when truncating examples.

Returns

Dataset or DatasetDict

The dataset with truncated sequences.

Truncate sequences in a dataset to a specified max_length.

Example:

>>> from datasets import Dataset

>>> examples = {
...     "input_ids": [[1, 2, 3], [4, 5, 6, 7], [8]],
...     "attention_mask": [[0, 1, 1], [0, 0, 1, 1], [1]],
... }
>>> dataset = Dataset.from_dict(examples)
>>> truncated_dataset = truncate_dataset(dataset, max_length=2)
>>> truncated_dataset[:]
{'input_ids': [[1, 2], [4, 5], [8]],
 'attention_mask': [[0, 1], [0, 0], [1]]}

Update on GitHub

TRL

Data Utilities

prepare_multimodal_messages

trl.prepare_multimodal_messages

prepare_multimodal_messages_vllm

trl.prepare_multimodal_messages_vllm

is_conversational

trl.is_conversational

is_conversational_from_value

trl.is_conversational_from_value

apply_chat_template

trl.apply_chat_template

maybe_apply_chat_template

trl.maybe_apply_chat_template

maybe_convert_to_chatml

trl.maybe_convert_to_chatml

extract_prompt

trl.extract_prompt

maybe_extract_prompt

trl.maybe_extract_prompt

unpair_preference_dataset

trl.unpair_preference_dataset

maybe_unpair_preference_dataset

trl.maybe_unpair_preference_dataset

pack_dataset

trl.pack_dataset

truncate_dataset

trl.truncate_dataset