Skip to content

Commit

Permalink
docs: Fix a typo in changelog regarding packing in SFT (#424)
Browse files Browse the repository at this point in the history
Signed-off-by: ashors1 <ashors@nvidia.com>
Co-authored-by: Terry Kong <terryk@nvidia.com>
  • Loading branch information
ashors1 and terrykong authored Dec 4, 2024
1 parent 5d4b2a7 commit c69b872
Showing 1 changed file with 2 additions and 2 deletions.
4 changes: 2 additions & 2 deletions CHANGELOG.md
Original file line number Diff line number Diff line change
Expand Up @@ -55,7 +55,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/)

### New Features and Optimizations
- Implement Kahneman-Tversky Optimization (KTO).
- Sequence packing is now supported when running SFT with SFTChatDataset.
- Sequence packing is now supported when running SFT with prompt-response datasets.

### Breaking Changes

Expand All @@ -75,7 +75,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/)
- Critic and Reward Model server refactored. Now the reward model will have a flag called `model.forward_micro_batch_size` which determines the micro batch size on which it runs inferences. This can be higher than the training micro batch size since during inference, we have less memory pressure.
- In the critic and reward model server, it is now possible to specify `inference_micro_batch_size` as a list. This allows us to provide more information to PyTriton regarding the preferred batch sizes for inference.
- It is no longer a requirement to specify `num_rollout_samples` to be a multiple of `inference_micro_batch_size * dp size` in PPO.
- Sequence packing is now supported when running SFT with SFTChatDataset.
- Sequence packing is now supported when running SFT with prompt-response datasets.
- Add online rejection sampling algorithm.

### Breaking Changes
Expand Down

0 comments on commit c69b872

Please sign in to comment.