Add test for SDPAToPagedAttention #12
Closed
Conversation
This file contains 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
…penvinotoolkit#308) ## Type of Change - The PR fixes an issue mentioned in openvinotoolkit#254. There was a fix proposed in openvinotoolkit#246, but this was only applied the cpp part and not the llm_bench/python. As we speak, with current code version, the llama2 conversion fails with same error when using llm_bench/python. - This PR introduces same fix as openvinotoolkit#246 for llm_bench/python too + fixes some missing documentation in README and typos ## Description - Same as above ## Expected Behavior & Potential Risk - N/A ## How has this PR been tested? - N/A ## Dependency Change? - N/A --------- Co-authored-by: Anas Ahouzi <[email protected]>
- [x] rely on optimum configs for export - [x] rely on optimum configs for weights compression - openvinotoolkit#218
Data-aware compression with updated configurations.
Bumps [diffusers](https://github.com/huggingface/diffusers) from 0.26.3 to 0.27.0. - [Release notes](https://github.com/huggingface/diffusers/releases) - [Commits](huggingface/diffusers@v0.26.3...v0.27.0) --- updated-dependencies: - dependency-name: diffusers dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <[email protected]>
Bumps [diffusers](https://github.com/huggingface/diffusers) from 0.26.3 to 0.27.0. - [Release notes](https://github.com/huggingface/diffusers/releases) - [Commits](huggingface/diffusers@v0.26.3...v0.27.0) --- updated-dependencies: - dependency-name: diffusers dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <[email protected]>
…search_with_hf
…t/pip/image_generation/stable_diffusion_1_5/cpp/scripts/diffusers-0.27.0 Bump diffusers from 0.26.3 to 0.27.0 in /image_generation/stable_diffusion_1_5/cpp/scripts
…t/pip/image_generation/lcm_dreamshaper_v7/cpp/scripts/diffusers-0.27.0 Bump diffusers from 0.26.3 to 0.27.0 in /image_generation/lcm_dreamshaper_v7/cpp/scripts
@p-wysocki @pavel-esir this fixes openvinotoolkit#302 , tested on local , added github actions workflow and updated the models list in README.md. --------- Co-authored-by: Pavel Esir <[email protected]>
` Skipping line in requirement file [openvino.genai/llm_bench/python/requirements.txt] because it's not clear what it would install: git+https://github.com/huggingface/optimum-intel.git@552de65a9c5f7fa1a2f0ce6859ebdeedaeaabe53 (add #egg=PackageName to the URL to avoid this warning) Skipping line in requirement file [openvino.genai/llm_bench/python/requirements.txt] because it's not clear what it would install: git+https://github.com/openvinotoolkit/nncf.git (add #egg=PackageName to the URL to avoid this warning) ` Signed-off-by: Peter Chen <[email protected]>
…issed add submodule update into sd installation instructions
I've verified support of Qwen1.5-7B by OpenVINO and then added it to the github workflow and readme.md ``` (base) root@8tvt:~/openvino.genai/llm_bench/python# ../../text_generation/causal_lm/cpp/build/greedy_causal_lm qwen/pytorch/dldt/FP32/ "Why is the Sun yellow?" The Sun does not actually appear yellow to us when we look at it. In fact, it appears white because it emits light across a wide range of wavelengths, including all the colors of the visible spectrum. When this light reaches our eyes, our eyes combine the different colors to create the perception of white. ``` --------- Co-authored-by: Pavel Esir <[email protected]>
Ticket: 140109
New shape approach
CVS-133717 1. if has option -ic, output token size is same as infer count 2. if without option -ic, output token size is generated by default according to the model. 3. remove the default output limit of 512 tokens. 4. if set env LOGLEVEL=DEBUG, will print latency of all tokens. examples: set env LOGLEVEL=DEBUG [bloomz-560m-without-ic.txt](https://github.com/openvinotoolkit/openvino.genai/files/15245407/bloomz-560m-without-ic.txt) [bloomz-560m-ic-1024.txt](https://github.com/openvinotoolkit/openvino.genai/files/15245409/bloomz-560m-ic-1024.txt) [llama-2-7b-chat-without-ic-.txt](https://github.com/openvinotoolkit/openvino.genai/files/15245412/llama-2-7b-chat-without-ic-.txt) [llama-2-7b-chat-without-ic-.txt](https://github.com/openvinotoolkit/openvino.genai/files/15245415/llama-2-7b-chat-without-ic-.txt) --------- Co-authored-by: Chen Peter <[email protected]>
Model lists in Python tests
Changed block_size according to latest CPU changes
Removed not suitable models from models list
Add a test to check if SDPAToPagedAttention transformation was performed Signed-off-by: Andrii Staikov <[email protected]>
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.
Add test for SDPAToPagedAttention
Add a test to check if SDPAToPagedAttention transformation
was performed
Signed-off-by: Andrii Staikov [email protected]