Add experimental support of fused decoder layer for llama2 #11768
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
plusbang
force-pushed
the
add-llama-fuseddecoder
branch
from
df4be6f
to
630dd23
Compare
yangw1234
reviewed
Aug 12, 2024
yangw1234
reviewed
Aug 12, 2024
There was a problem hiding this comment.
I think the changes in pipeline_parallel.py, if cleaned up a bit , should not affect gpu code. Maybe we should try to merge them later.
yangw1234
approved these changes
Aug 12, 2024
yangw1234
reviewed
Aug 12, 2024
| pip install --pre --upgrade ipex-llm[all] | ||
| pip install --pre --upgrade bigdl-core-npu | ||
|
|
||
| pip install transformers==4.40 |
There was a problem hiding this comment.
I am experimenting on transformers==3.39.3, have we verified 4.40?
There was a problem hiding this comment.
I am experimenting on
transformers==3.39.3, have we verified 4.40?
yes, I have also verified transformers==4.40.
Have added optimize_llm_post to wrap lm_head related logic etc. And considering we may need additional process to address prefill, maybe we could keep separating npu and gpu PP temporarilly.
|
Merge it first, please let me know if any suggestion or comment. |
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Outdated suggestions cannot be applied.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
Refactor #11716
This experimental support is only verified on llama2-7b currently.
To avoid influence on GPU support and NPU support for other models (llama3, qwen2, etc), add
DynamicFusedNormalCache,llama_fused_model_forwardandpipeline_parallelintransformers/npu_models/. Then only specifypipeline_parallel_stages > 1will use related logic.Add experimental example and readme
4. How to test?
1234). And paste your action link here once it has been successfully finished.https://github.com/intel-analytics/ipex-llm-workflow/actions/runs/10362835964