Add CrossAttention operator #14146
Merged
Add CrossAttention operator #14146
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
tianleiwu
changed the title
wangyems
reviewed
Jan 6, 2023
| "value", | ||
| "Value with shape (batch_size, kv_sequence_length, v_hidden_size)", | ||
| "T") | ||
| .Input(3, |
Contributor
There was a problem hiding this comment.
optional? (if combine kv in future)
Contributor
Author
There was a problem hiding this comment.
Packed kv will be supported later. Will update the interface accordingly at that time.
wangyems
reviewed
Jan 6, 2023
| // Input 6 (value) has shape (batch_size, kv_sequence_length, v_hidden_size) | ||
| // | ||
| // Input 0 has 3D shape (batch_size, sequence_length, input_hidden_size) | ||
| // INput 1 has 2D shape (input_hidden_size, hidden_size + hidden_size + v_hidden_size) |
Contributor
Author
There was a problem hiding this comment.
will fix it in next PR.
wangyems
approved these changes
Jan 6, 2023
tianleiwu
added a commit
that referenced
this pull request
Jan 10, 2023
We remove key and value inputs in #14146, need update the convert_generation as well.
hanbitmyths
pushed a commit
that referenced
this pull request
Apr 19, 2023
### Description This PR contains fusion-level and kernel-level optimizations for [OpenAI's Whisper](https://github.com/openai/whisper). Some of the added optimizations include: - Pruning of duplicate/unnecessary inputs and outputs - Fusion support for Whisper models with or without these inputs/outputs (e.g. with these inputs/outputs if exporting with an older official Optimum version, without these inputs/outputs if exporting with Optimum from source) - Attention fusions - For Whisper's encoder and decoder - Modified symbolic shape inference for present output when no past input exists (for decoder) - Multi-head attention fusions - For Whisper's decoder and decoder with past - Packed MatMul for the 3 MatMuls excluded in multi-head attention fusion - Attention kernel changes - CPU: - Different Q and KV sequence lengths - Parallel memset for large sequence lengths - Convert broadcast add after MatMul of Q and K (add_qk) to element-wise add - Separate present key-value output into present key and present value (for multi-head attention spec) - CUDA: - Use memory efficient attention compute kernel with present state (for decoder) - Multi-head attention kernel changes - CPU: - Introduction of multi-head attention CPU kernel (previously did not exist) - Use AddBiasReshape instead of AddBiasTranspose when sequence length = 1 (for decoder with past) - Different Q, K, V input shapes - Pass past key and past value directly as key and value - CUDA: - Use memory efficient attention compute kernel with past and/or present state (for decoder with past) ### Usage To use the optimizations, run the ORT transformer optimizer script as follows: ``` $ cd onnxruntime/onnxruntime/python/tools/transformers/ $ python3 optimizer.py --input <filename>.onnx --output <filename>.onnx --model_type bart --num_heads <number of attention heads, depends on the size of the whisper model used> --hidden_size <attention hidden size, depends on the size of the whisper model used> --use_external_data_format --use_multi_head_attention ``` Once optimized, here's an example of how to run Whisper with [Hugging Face's Optimum](https://github.com/huggingface/optimum): ``` from transformers.onnx.utils import get_preprocessor from optimum.onnxruntime import ORTModelForSpeechSeq2Seq from optimum.pipelines import pipeline as ort_pipeline import whisper # Installed from OpenAI's repo - setup instructions at https://github.com/openai/whisper/ directory = './whisper_opt' # Where the optimized ONNX models are located model_name = 'openai/whisper-tiny' device = 'cpu' # Get pipeline processor = get_preprocessor(model_name) model = ORTModelForSpeechSeq2Seq.from_pretrained( directory, use_io_binding=(device == 'cuda'), provider='CPUExecutionProvider', ).to(device) pipe = ort_pipeline( "automatic-speech-recognition", model=model, tokenizer=processor.tokenizer, feature_extractor=processor.feature_extractor, device=(-1 if device == 'cpu' else 0), ) # Load audio file and run pipeline audio = whisper.load_audio('tests/jfk.flac') audio = whisper.pad_or_trim(audio) outputs = pipe([audio]) print(outputs) ``` Note: In order to use these changes with Optimum, it is recommended to use Optimum from source to have the following changes: - huggingface/optimum#872 - huggingface/optimum#920 ### Motivation and Context This PR helps the following issues: - #15100 - #15235 - huggingface/optimum#869 (work in progress) This PR can be used with the other currently merged Whisper PRs: - #15247 - #15339 - #15362 - #15365 - #15427 This PR uses changes from the following merged PRs: - #14198 - #14146 - #14201 - #14928 (this introduced the new multi-head attention spec)
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.
Description
Move separated Q, K and V (without input projection) from Attention to a new operator CrossAttention.
The Attention operator is hard to maintain when we need support with and without input projection in one class. Add a new operator according to feedback.
Some change might need in the future, but not in this PR:
(1) bias could be optional (We will not proceed that route unless experiments show that fusing Add bias with MatMul instead of this op could improve performance).
(2) support packed KV. There are two ways to support it: when key and value are same Tensor, they are packed; or we can make value as optional, and use packed mode when value is empty and the key has packed K/V.
(3) support cached key and value, and other (like relative position bias), or more attention mask format. They can be added easily without breaking backward compatible.
(4) ROCm/CPU implementation of this op.
Motivation and Context