FIX Beam search w/ mixed adapter batches & encoder#2921
Merged
BenjaminBossan merged 1 commit intohuggingface:mainfrom Nov 20, 2025
Merged
Conversation
When using mixed adapter batches (i.e. using different LoRA adapters in the same batch), users have to pass adapter_names. When simultaneously using beam search, these adapter names have to be extended by the number of beams. For encoder-decoder models, even when applying beam search, the encoder part of the model should, however, not use the extended adapter_names. This is because the encoder still uses the original, non-extended samples. The need for this used to be checked by calling model.get_encoder(). However, with transformers v5, every PretrainedModel will have a get_encoder method. The new convention is that it will return self if there is no encoder. This is now what's being checked. huggingface/transformers#42156 Note that said PR contains a small bug that leads to self not always being returned. Therefore, for the full fix of the issue on transformers main, we also need to await this PR: huggingface/transformers#42295
|
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. |
githubnemo
approved these changes
Nov 20, 2025
BenjaminBossan
deleted the
fix-beam-search-mixed-adapter-batch-encoder
branch
Conzel
pushed a commit
to Conzel/peft
that referenced
this pull request
Nov 25, 2025
When using mixed adapter batches (i.e. using different LoRA adapters in the same batch), users have to pass adapter_names. When simultaneously using beam search, these adapter names have to be extended by the number of beams. For encoder-decoder models, even when applying beam search, the encoder part of the model should, however, not use the extended adapter_names. This is because the encoder still uses the original, non-extended samples. The need for this used to be checked by calling model.get_encoder(). However, with transformers v5, every PretrainedModel will have a get_encoder method. The new convention is that it will return self if there is no encoder. This is now what's being checked. huggingface/transformers#42156 Note that said PR contains a small bug that leads to self not always being returned. Therefore, for the full fix of the issue on transformers main, we also need to await this PR: huggingface/transformers#42295
Arlaz
pushed a commit
to obvious-research/peft
that referenced
this pull request
Jan 9, 2026
When using mixed adapter batches (i.e. using different LoRA adapters in the same batch), users have to pass adapter_names. When simultaneously using beam search, these adapter names have to be extended by the number of beams. For encoder-decoder models, even when applying beam search, the encoder part of the model should, however, not use the extended adapter_names. This is because the encoder still uses the original, non-extended samples. The need for this used to be checked by calling model.get_encoder(). However, with transformers v5, every PretrainedModel will have a get_encoder method. The new convention is that it will return self if there is no encoder. This is now what's being checked. huggingface/transformers#42156 Note that said PR contains a small bug that leads to self not always being returned. Therefore, for the full fix of the issue on transformers main, we also need to await this PR: huggingface/transformers#42295
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
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.
When using mixed adapter batches (i.e. using different LoRA adapters in the same batch), users have to pass
adapter_names(one name per sample). When simultaneously using beam search, these adapter names have to be extended by the number of beams. For encoder-decoder models, even when applying beam search, the encoder part of the model should, however, not use the extendedadapter_names. This is because the encoder still uses the original, non-extended samples.The need for this used to be checked by calling
model.get_encoder(). However, with transformers v5, everyPretrainedModelwill have aget_encodermethod. The new convention is that it will returnselfif there is no encoder. This is now what's being checked.Note that the transformers PR contains a small bug that leads to
selfnot always being returned. Therefore, for the full fix of the issue on transformers main, we also need to await this PR:huggingface/transformers#42295
I checked the failing tests with that PR applied and they pass.