Deprecate unused dataset_formatting module #4242
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
qgallouedec
reviewed
Oct 9, 2025
Per reviewer feedback, adding FutureWarning to all functions in dataset_formatting module instead of immediate removal. Functions will be removed in TRL 0.27. - Added warnings.warn() to get_formatting_func_from_dataset() - Added warnings.warn() to conversations_formatting_function() - Added warnings.warn() to instructions_formatting_function() - Updated docstrings with deprecation notices - Added pytest.mark.filterwarnings to test class to suppress expected warnings
behroozazarkhalili
force-pushed
the
remove-dataset-formatting-module
branch
from
1baa79b to
eda2f04
Compare
|
✅ Added deprecation warnings instead of immediate removal. Changes made:
All functions now warn users to use |
Resolve qgallouedec's review comment by: - Deleting docs/source/detoxifying_a_lm.md (obsolete toxicity documentation) - Removing reference from _toctree.yml - Removing research_projects references from example_overview.md - Removing stack_llama script references from peft_integration.md All removed documentation referenced deleted research_projects scripts that no longer exist in the repository.
This reverts commit 24cb61a.
|
Can you please rename the PR "deprecate" |
behroozazarkhalili
changed the title
Done 👍 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Analysis:
trl.extras.dataset_formatting📁 Current State
File:
trl/extras/dataset_formatting.pyFunctions:
get_formatting_func_from_dataset()- Auto-detects dataset formatconversations_formatting_function()- Formats ChatML-style conversationsinstructions_formatting_function()- Formats instruction-completion pairs🔍 Usage Analysis
Internal Usage:
trl/extras/__init__.pytrl/__init__.py(public API)SFTTrainer,GKDTrainer, etc.)tests/test_dataset_formatting.pyExternal Usage Risk:
from trl.extras.dataset_formatting import get_formatting_func_from_dataset📜 History
Added: PR #1208 (ChatML support)
Last Modified: July 2025, PR #3704 - removed
ConstantLengthDatasetdependencyDeprecation Status: ❌ No warnings currently in place
🤔 Why It's Obsolete
tokenizer.apply_chat_template()SFTTraineraccepts user-providedformatting_func, not auto-detected onesChanges
trl/extras/dataset_formatting.pytests/test_dataset_formatting.py