Do not initialise ref model when peft is used #224
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
When running the DPO script with a PEFT configuration, e.g.: ``` ACCELERATE_LOG_LEVEL=info accelerate launch --config_file recipes/accelerate_configs/ddp.yaml --num_processes=1 scripts/dpo.py --config recipes/zephyr-7b-beta/dpo/config_qlora.yaml ``` the following error is raised: ```ValueError: You passed both a ref_model and a peft_config. For training PEFT adapters with DPO there is no need to pass a reference model. Please pass `ref_model=None` in case you want to train PEFT adapters, or pass a ref_model with `force_use_ref_model=True` in DPOTrainer's init. if you want to use a different ref_model.``` This happens because the script tries to initialize a reference model even when LoRA/PEFT is enabled. This PR updates the logic to skip reference model initialization when PEFT adapters are used. I am referring to the similar logic that is used in the trl example dpo script: https://github.com/huggingface/trl/blob/main/trl/scripts/dpo.py#L109
|
Hey @lewtun could you please have a look at this when you have time? My team would benefit from this fix :) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.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.
When running the DPO script with a PEFT configuration, e.g.:
the following error is raised:
ValueError: You passed both a ref_model and a peft_config. For training PEFT adapters with DPO there is no need to pass a reference model. Please pass `ref_model=None` in case you want to train PEFT adapters, or pass a ref_model with `force_use_ref_model=True` in DPOTrainer's init. if you want to use a different ref_model.This happens because the script tries to initialize a reference model even when LoRA/PEFT is enabled.
This PR updates the logic to skip reference model initialization when PEFT adapters are used.
I am referring to the similar logic that is used in the trl example dpo script:
https://github.com/huggingface/trl/blob/main/trl/scripts/dpo.py#L109