[None][feat] VeRL/TRTLLM prototype #9133
Conversation
| self.app.add_api_route("/resume_memory", | ||
| self.resume_memory, | ||
| methods=["POST"]) | ||
| self.app.add_api_route("/update_weights", |
There was a problem hiding this comment.
Note: weight update is in worker extension instead of in LLM() class. So the end point part needs refactored as well.
| async def async_worker(rank, world_size, model_dir, tensor_parallel_size, use_fsdp): | ||
| #os.environ["CUDA_VISIBLE_DEVICES"] = "0,1,2,3,4,5,6,7" | ||
| #os.environ["TRTLLM_RAY_BUNDLE_INDICES"] = "1,2,3,4,5,6,7" | ||
| os.environ["CUDA_VISIBLE_DEVICES"] = "0,1" |
There was a problem hiding this comment.
We don't need to use CUDA_VISIBLE_DEVICES if you're already using TRTLLM_RAY_BUNDLE_INDICES. RayExecutor needs global view of GPUs, otherwise it'd have issues when doing placement via bundle index.
| def __init__(self, *args, **kwargs): | ||
| super().__init__(*args, **kwargs) | ||
|
|
||
| async def async_init_phase(self): |
There was a problem hiding this comment.
Notes from syncing w/ Liwei:
VeRL needs async init for TRTLLM's LLM(), but PYthon has a limitation where init must be sync. So Liwei separate the async part out here.
|
|
||
| world_size = args.tensor_parallel_size | ||
| master_port = get_free_port() | ||
| mp.spawn(worker, args=(world_size, master_port, args.model_dir, args.tensor_parallel_size, args.use_fsdp), nprocs=world_size, join=True) |
| @@ -0,0 +1,647 @@ | |||
| import argparse | |||
| import torch | |||
There was a problem hiding this comment.
Note: Should be updated to use TRTLLMHttpServer from verl-integ/verl/workers/rollout/trtllm_rollout/trtllm_async_server.py
| @field_validator("gpus_per_node", mode='before') | ||
| @classmethod | ||
| def validate_gpus_per_node(cls, v, info): | ||
| if os.getenv("RAY_LOCAL_WORLD_SIZE") is not None: |
There was a problem hiding this comment.
Note: According to Liwei, this is likely obsolete.
trtllm works in async env trtllm works in sync and async env ray + updatew works rebase to the updated verl server mode still cherry pick still cherry pick still cherry pick integrated http interface hang at RyExecutor create workers ray.remote clean code use tensorrt_llm.rlhf_utils
hchings
changed the title
hchings
changed the title
one pg works from test, with 2 in 8 bundle indices update weights should work. move to eos trtllm (ray) + verl first works trtllm (ray) + verl first works; bundled pickle
2 participants
Description
Test Coverage
PR Checklist
Please review the following before submitting your PR:
PR description clearly explains what and why. If using CodeRabbit's summary, please make sure it makes sense.
PR Follows TRT-LLM CODING GUIDELINES to the best of your knowledge.
Test cases are provided for new code paths (see test instructions)
Any new dependencies have been scanned for license and vulnerabilities
CODEOWNERS updated if ownership changes
Documentation updated as needed
Update tava architecture diagram if there is a significant design change in PR.
The reviewers assigned automatically/manually are appropriate for the PR.
Please check this after reviewing the above items as appropriate for this PR.
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.
Details
run [--reuse-test (optional)pipeline-id --disable-fail-fast --skip-test --stage-list "A10-PyTorch-1, xxx" --gpu-type "A30, H100_PCIe" --test-backend "pytorch, cpp" --add-multi-gpu-test --only-multi-gpu-test --disable-multi-gpu-test --post-merge --extra-stage "H100_PCIe-TensorRT-Post-Merge-1, xxx" --detailed-log --debug(experimental)]Launch build/test pipelines. All previously running jobs will be killed.
--reuse-test (optional)pipeline-id(OPTIONAL) : Allow the new pipeline to reuse build artifacts and skip successful test stages from a specified pipeline or the last pipeline if no pipeline-id is indicated. If the Git commit ID has changed, this option will be always ignored. The DEFAULT behavior of the bot is to reuse build artifacts and successful test results from the last pipeline.--disable-reuse-test(OPTIONAL) : Explicitly prevent the pipeline from reusing build artifacts and skipping successful test stages from a previous pipeline. Ensure that all builds and tests are run regardless of previous successes.--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-PyTorch-1, xxx"(OPTIONAL) : Only run the specified test stages. Examples: "A10-PyTorch-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.--test-backend "pytorch, cpp"(OPTIONAL) : Skip test stages which don't match the specified backends. Only support [pytorch, cpp, tensorrt, triton]. Examples: "pytorch, cpp" (does not run test stages with tensorrt or triton backend). Note: Does NOT update GitHub pipeline 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 in addition to running 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-TensorRT-Post-Merge-1, xxx"(OPTIONAL) : Run the ordinary L0 pre-merge pipeline and specified test stages. Examples: --extra-stage "H100_PCIe-TensorRT-Post-Merge-1, xxx".--detailed-log(OPTIONAL) : Enable flushing out all logs to the Jenkins console. This will significantly increase the log volume and may slow down the job.--debug(OPTIONAL) : Experimental feature. Enable access to the CI container for debugging purpose. Note: Specify exactly one stage in thestage-listparameter to access the appropriate container environment. Note: Does NOT update GitHub check status.For guidance on mapping tests to stage names, see
docs/source/reference/ci-overview.mdand the
scripts/test_to_stage_mapping.pyhelper.kill
killKill all running builds associated with pull request.
skip
skip --comment COMMENTSkip 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-pipelineReuse 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.