♻️ Fix caching in SFT #2945
Merged
♻️ Fix caching in SFT #2945
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
kashif
approved these changes
Feb 24, 2025
|
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. |
edbeeching
approved these changes
Feb 24, 2025
There was a problem hiding this comment.
LGTM. Out of curiousity. If I run two experiments with different max_length (and hence different packed dataset sizes), will both packed datasets be cached independently, so future exps will reuse the correct cached dataset, or does caching one overwrite the other?
Yes! from trl import SFTConfig, SFTTrainer
from datasets import load_dataset
dataset = load_dataset("trl-lib/Capybara", split="train[:10%]")
# Processes the dataset
training_args = SFTConfig(output_dir="Qwen/Qwen2.5-0.5B-SFT", max_length=128, packing=True)
trainer = SFTTrainer(
args=training_args,
model="Qwen/Qwen2.5-0.5B",
train_dataset=dataset,
)
# Processes the dataset as well
training_args = SFTConfig(output_dir="Qwen/Qwen2.5-0.5B-SFT", max_length=256, packing=True)
trainer = SFTTrainer(
args=training_args,
model="Qwen/Qwen2.5-0.5B",
train_dataset=dataset,
)
# Uses the cache!
training_args = SFTConfig(output_dir="Qwen/Qwen2.5-0.5B-SFT", max_length=128, packing=True)
trainer = SFTTrainer(
args=training_args,
model="Qwen/Qwen2.5-0.5B",
train_dataset=dataset,
) |
qgallouedec
added a commit
that referenced
this pull request
Feb 25, 2025
jhinpan
pushed a commit
to jhinpan/trl-jin
that referenced
this pull request
Mar 12, 2025
yxliu-TAMU
pushed a commit
to mincheolseong/ECEN743-GRPO-Project-Proposal
that referenced
this pull request
Apr 20, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
What does this PR do?
Fixes # (issue)
Before submitting
Pull Request section?
to it if that's the case.
documentation guidelines.
Who can review?
Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.