Skip to content

Commit

Permalink
[RoBERTa-based] Add support for sdpa (#30510)
Browse files Browse the repository at this point in the history 
* Adding SDPA support for RoBERTa-based models

* add not is_cross_attention

* fix copies

* fix test

* add minimal test for camembert and xlm_roberta as their test class does not inherit from ModelTesterMixin

* address some review comments

* use copied from

* style

* consistency

* fix lists

---------

Co-authored-by: fxmarty <[email protected]>
Co-authored-by: Arthur <[email protected]>
  • Loading branch information
@hackyon @fxmarty @ArthurZucker
3 people authored Aug 28, 2024
1 parent e0b87b0 commit f1a385b
Show file tree
Hide file tree
Showing 11 changed files with 828 additions and 100 deletions.
27 changes: 20 additions & 7 deletions docs/source/en/perf_infer_gpu_one.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -74,7 +74,6 @@ FlashAttention-2 is currently supported for the following architectures:
* [OPT](https://huggingface.co/docs/transformers/model_doc/opt#transformers.OPTModel)
* [Phi](https://huggingface.co/docs/transformers/model_doc/phi#transformers.PhiModel)
* [Phi3](https://huggingface.co/docs/transformers/model_doc/phi3#transformers.Phi3Model)
* [SigLIP](https://huggingface.co/docs/transformers/model_doc/siglip)
* [StableLm](https://huggingface.co/docs/transformers/model_doc/stablelm#transformers.StableLmModel)
* [Starcoder2](https://huggingface.co/docs/transformers/model_doc/starcoder2#transformers.Starcoder2Model)
* [Qwen2](https://huggingface.co/docs/transformers/model_doc/qwen2#transformers.Qwen2Model)
Expand All @@ -86,6 +85,7 @@ FlashAttention-2 is currently supported for the following architectures:
* [Hubert](https://huggingface.co/docs/transformers/model_doc/hubert#transformers.HubertModel)
* [data2vec_audio](https://huggingface.co/docs/transformers/main/en/model_doc/data2vec#transformers.Data2VecAudioModel)
* [Sew](https://huggingface.co/docs/transformers/main/en/model_doc/sew#transformers.SEWModel)
* [SigLIP](https://huggingface.co/docs/transformers/model_doc/siglip)
* [UniSpeech](https://huggingface.co/docs/transformers/v4.39.3/en/model_doc/unispeech#transformers.UniSpeechModel)
* [unispeech_sat](https://huggingface.co/docs/transformers/v4.39.3/en/model_doc/unispeech-sat#transformers.UniSpeechSatModel)

Expand Down Expand Up @@ -204,9 +204,11 @@ For now, Transformers supports SDPA inference and training for the following arc
* [Audio Spectrogram Transformer](https://huggingface.co/docs/transformers/model_doc/audio-spectrogram-transformer#transformers.ASTModel)
* [Bart](https://huggingface.co/docs/transformers/model_doc/bart#transformers.BartModel)
* [Bert](https://huggingface.co/docs/transformers/model_doc/bert#transformers.BertModel)
* [CamemBERT](https://huggingface.co/docs/transformers/model_doc/camembert#transformers.CamembertModel)
* [Chameleon](https://huggingface.co/docs/transformers/model_doc/chameleon#transformers.Chameleon)
* [CLIP](https://huggingface.co/docs/transformers/model_doc/clip#transformers.CLIPModel)
* [Cohere](https://huggingface.co/docs/transformers/model_doc/cohere#transformers.CohereModel)
* [data2vec_audio](https://huggingface.co/docs/transformers/main/en/model_doc/data2vec#transformers.Data2VecAudioModel)
* [Dbrx](https://huggingface.co/docs/transformers/model_doc/dbrx#transformers.DbrxModel)
* [DeiT](https://huggingface.co/docs/transformers/model_doc/deit#transformers.DeiTModel)
* [Dpr](https://huggingface.co/docs/transformers/model_doc/dpr#transformers.DprReader)
Expand All @@ -216,10 +218,16 @@ For now, Transformers supports SDPA inference and training for the following arc
* [GPT2](https://huggingface.co/docs/transformers/model_doc/gpt2)
* [GPTBigCode](https://huggingface.co/docs/transformers/model_doc/gpt_bigcode#transformers.GPTBigCodeModel)
* [GPTNeoX](https://huggingface.co/docs/transformers/model_doc/gpt_neox#transformers.GPTNeoXModel)
* [Hubert](https://huggingface.co/docs/transformers/model_doc/hubert#transformers.HubertModel)
* [Idefics](https://huggingface.co/docs/transformers/model_doc/idefics#transformers.IdeficsModel)
* [Granite](https://huggingface.co/docs/transformers/model_doc/granite#transformers.GraniteModel)
* [JetMoe](https://huggingface.co/docs/transformers/model_doc/jetmoe#transformers.JetMoeModel)
* [Jamba](https://huggingface.co/docs/transformers/model_doc/jamba#transformers.JambaModel)
* [Llama](https://huggingface.co/docs/transformers/model_doc/llama#transformers.LlamaModel)
* [Mistral](https://huggingface.co/docs/transformers/model_doc/mistral#transformers.MistralModel)
* [Mixtral](https://huggingface.co/docs/transformers/model_doc/mixtral#transformers.MixtralModel)
* [Musicgen](https://huggingface.co/docs/transformers/model_doc/musicgen#transformers.MusicgenModel)
* [MusicGen Melody](https://huggingface.co/docs/transformers/model_doc/musicgen_melody#transformers.MusicgenMelodyModel)
* [OLMo](https://huggingface.co/docs/transformers/model_doc/olmo#transformers.OlmoModel)
* [PaliGemma](https://huggingface.co/docs/transformers/model_doc/paligemma#transformers.PaliGemmaForConditionalGeneration)
* [Phi](https://huggingface.co/docs/transformers/model_doc/phi#transformers.PhiModel)
Expand All @@ -233,6 +241,14 @@ For now, Transformers supports SDPA inference and training for the following arc
* [Qwen2](https://huggingface.co/docs/transformers/model_doc/qwen2#transformers.Qwen2Model)
* [Qwen2Audio](https://huggingface.co/docs/transformers/model_doc/qwen2_audio#transformers.Qwen2AudioEncoder)
* [Qwen2MoE](https://huggingface.co/docs/transformers/model_doc/qwen2_moe#transformers.Qwen2MoeModel)
* [RoBERTa](https://huggingface.co/docs/transformers/model_doc/roberta#transformers.RobertaModel)
* [Sew](https://huggingface.co/docs/transformers/main/en/model_doc/sew#transformers.SEWModel)
* [SigLIP](https://huggingface.co/docs/transformers/model_doc/siglip)
* [StableLm](https://huggingface.co/docs/transformers/model_doc/stablelm#transformers.StableLmModel)
* [Starcoder2](https://huggingface.co/docs/transformers/model_doc/starcoder2#transformers.Starcoder2Model)
* [UniSpeech](https://huggingface.co/docs/transformers/v4.39.3/en/model_doc/unispeech#transformers.UniSpeechModel)
* [unispeech_sat](https://huggingface.co/docs/transformers/v4.39.3/en/model_doc/unispeech-sat#transformers.UniSpeechSatModel)
* [RoBERTa](https://huggingface.co/docs/transformers/model_doc/roberta#transformers.RobertaModel)
* [Qwen2VL](https://huggingface.co/docs/transformers/model_doc/qwen2_vl#transformers.Qwen2VLModel)
* [Musicgen](https://huggingface.co/docs/transformers/model_doc/musicgen#transformers.MusicgenModel)
* [MusicGen Melody](https://huggingface.co/docs/transformers/model_doc/musicgen_melody#transformers.MusicgenMelodyModel)
Expand All @@ -243,12 +259,9 @@ For now, Transformers supports SDPA inference and training for the following arc
* [ViTMSN](https://huggingface.co/docs/transformers/model_doc/vit_msn#transformers.ViTMSNModel)
* [VideoMAE](https://huggingface.co/docs/transformers/model_doc/videomae#transformers.VideoMAEModell)
* [wav2vec2](https://huggingface.co/docs/transformers/model_doc/wav2vec2#transformers.Wav2Vec2Model)
* [Hubert](https://huggingface.co/docs/transformers/model_doc/hubert#transformers.HubertModel)
* [data2vec_audio](https://huggingface.co/docs/transformers/main/en/model_doc/data2vec#transformers.Data2VecAudioModel)
* [SigLIP](https://huggingface.co/docs/transformers/model_doc/siglip)
* [Sew](https://huggingface.co/docs/transformers/main/en/model_doc/sew#transformers.SEWModel)
* [UniSpeech](https://huggingface.co/docs/transformers/v4.39.3/en/model_doc/unispeech#transformers.UniSpeechModel)
* [unispeech_sat](https://huggingface.co/docs/transformers/v4.39.3/en/model_doc/unispeech-sat#transformers.UniSpeechSatModel)
* [Whisper](https://huggingface.co/docs/transformers/model_doc/whisper#transformers.WhisperModel)
* [XLM-RoBERTa](https://huggingface.co/docs/transformers/model_doc/xlm-roberta#transformers.XLMRobertaModel)
* [XLM-RoBERTa-XL](https://huggingface.co/docs/transformers/model_doc/xlm-roberta-xl#transformers.XLMRobertaXLModel)
* [YOLOS](https://huggingface.co/docs/transformers/model_doc/yolos#transformers.YolosModel)


Expand Down
2 changes: 1 addition & 1 deletion src/transformers/models/bridgetower/modeling_bridgetower.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -1063,7 +1063,7 @@ class PreTrainedModel
for layer, heads in heads_to_prune.items():
self.encoder.layer[layer].attention.prune_heads(heads)

# Copied from transformers.models.roberta.modeling_roberta.RobertaModel.forward
# Copied from transformers.models.clap.modeling_clap.ClapTextModel.forward
def forward(
self,
input_ids: Optional[torch.Tensor] = None,
Expand Down
Loading

0 comments on commit f1a385b

Please sign in to comment.