NVIDIA
[TRTLLM-5273]feat/Use full attention mask if Llama3 is used as encoder and fix EarlyStopDecoder unsqueeze bug
[TRTLLM-5273]feat/Use full attention mask if Llama3 is used as encoder and fix EarlyStopDecoder unsqueeze bug
Description
This PR adds in a flag for bidirectional_attention to modeling_llama.py. This is to support some customer use-cases for llama-based embedding models, see Jira TRTLLM-5273 for more details.
Additionally it fixes a bug in EarlyStopDecoder where 1d logits were being unsqueezed incorrectly, leading to incompatibility with LogitsStorage.
Lastly it hardcodes LlamaForTextEmbedding as a non-generation model for now so that the customer can make their own out-of-tree example with their custom model and still leverage EarlyStopDecoder. I am in discussions with Erin Ho and Freddy Qi about how we can eliminate the need to hardcode this in the future.
Things To Review
- Should this bidirectional attention be supported via flag like the current implementation or should it be a separate class?
GitHub Bot Help
/bot [-h] ['run', 'kill', 'skip', 'reuse-pipeline'] ...
Provide a user friendly way for developers to interact with a Jenkins server.
Run /bot [-h|--help] to print this help message.
See details below for each supported subcommand.
run [--disable-fail-fast --skip-test --stage-list "A10-1, xxx" --gpu-type "A30, H100_PCIe" --add-multi-gpu-test --only-multi-gpu-test --disable-multi-gpu-test --post-merge --extra-stage "H100_PCIe-[Post-Merge]-1, xxx"]
Launch build/test pipelines. All previously running jobs will be killed.
--disable-fail-fast (OPTIONAL) : Disable fail fast on build/tests/infra failures.
--skip-test (OPTIONAL) : Skip all test stages, but still run build stages, package stages and sanity check stages. Note: Does NOT update GitHub check status.
--stage-list "A10-1, xxx" (OPTIONAL) : Only run the specified test stages. Examples: "A10-1, xxx". Note: Does NOT update GitHub check status.
--gpu-type "A30, H100_PCIe" (OPTIONAL) : Only run the test stages on the specified GPU types. Examples: "A30, H100_PCIe". Note: Does NOT update GitHub check status.
--only-multi-gpu-test (OPTIONAL) : Only run the multi-GPU tests. Note: Does NOT update GitHub check status.
--disable-multi-gpu-test (OPTIONAL) : Disable the multi-GPU tests. Note: Does NOT update GitHub check status.
--add-multi-gpu-test (OPTIONAL) : Force run the multi-GPU tests. Will also run L0 pre-merge pipeline.
--post-merge (OPTIONAL) : Run the L0 post-merge pipeline instead of the ordinary L0 pre-merge pipeline.
--extra-stage "H100_PCIe-[Post-Merge]-1, xxx" (OPTIONAL) : Run the ordinary L0 pre-merge pipeline and specified test stages. Examples: --extra-stage "H100_PCIe-[Post-Merge]-1, xxx".
kill
kill
Kill all running builds associated with pull request.
skip
skip --comment COMMENT
Skip testing for latest commit on pull request. --comment "Reason for skipping build/test" is required. IMPORTANT NOTE: This is dangerous since lack of user care and validation can cause top of tree to break.
reuse-pipeline
reuse-pipeline
Reuse a previous pipeline to validate current commit. This action will also kill all currently running builds associated with the pull request. IMPORTANT NOTE: This is dangerous since lack of user care and validation can cause top of tree to break.
PR_Github #5279 [ run ] completed with state FAILURE
/LLM/main/L0_MergeRequest_PR pipeline #3857 completed with status: 'FAILURE'
PR_Github #5363 [ run ] completed with state SUCCESS
/LLM/main/L0_MergeRequest_PR pipeline #3916 completed with status: 'FAILURE'
PR_Github #5560 [ run ] completed with state SUCCESS
/LLM/main/L0_MergeRequest_PR pipeline #4057 completed with status: 'FAILURE'
PR_Github #5577 [ run ] completed with state SUCCESS
/LLM/main/L0_MergeRequest_PR pipeline #4067 completed with status: 'FAILURE'
PR_Github #5638 [ run ] completed with state FAILURE
/LLM/main/L0_MergeRequest_PR pipeline #4120 completed with status: 'FAILURE'
PR_Github #5771 [ run ] completed with state SUCCESS
/LLM/main/L0_MergeRequest_PR pipeline #4225 completed with status: 'SUCCESS'
