biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[V1] V1 Enablement Oracle (#13726)

Browse Source 
Signed-off-by: rshaw@neuralmagic.com <rshaw@neuralmagic.com>
Co-authored-by: rshaw@neuralmagic.com <rshaw@neuralmagic.com>
Co-authored-by: Nicolò Lucchesi <nlucches@redhat.com>
Co-authored-by: Tyler Michael Smith <tyler@neuralmagic.com>
Co-authored-by: Michael Goin <michael@neuralmagic.com>
This commit is contained in:
Robert Shaw
2025-03-15 01:02:20 -04:00
committed by GitHub
parent 8c0d15d5c5
commit d4d93db2c5
96 changed files with 1537 additions and 512 deletions
Download Patch File Download Diff File Expand all files Collapse all files

34
.buildkite/test-pipeline.yaml
View File

@@ -117,10 +117,10 @@ steps:
- pytest -v -s entrypoints/llm/test_lazy_outlines.py # it needs a clean process
- pytest -v -s entrypoints/llm/test_generate.py # it needs a clean process
- pytest -v -s entrypoints/llm/test_generate_multiple_loras.py # it needs a clean process
- pytest -v -s entrypoints/llm/test_guided_generate.py # it needs a clean process
- VLLM_USE_V1=0 pytest -v -s entrypoints/llm/test_guided_generate.py # it needs a clean process
- pytest -v -s entrypoints/openai --ignore=entrypoints/openai/test_oot_registration.py --ignore=entrypoints/openai/correctness/
- pytest -v -s entrypoints/test_chat_utils.py
- pytest -v -s entrypoints/offline_mode # Needs to avoid interference with other tests
- VLLM_USE_V1=0 pytest -v -s entrypoints/offline_mode # Needs to avoid interference with other tests
- label: Distributed Tests (4 GPUs) # 10min
working_dir: "/vllm-workspace/tests"
@@ -136,7 +136,7 @@ steps:
- examples/offline_inference/rlhf_colocate.py
- tests/examples/offline_inference/data_parallel.py
commands:
- VLLM_USE_V1=1 python3 ../examples/offline_inference/data_parallel.py
- python3 ../examples/offline_inference/data_parallel.py
- pytest -v -s distributed/test_utils.py
- pytest -v -s compile/test_basic_correctness.py
- pytest -v -s distributed/test_pynccl.py
@@ -197,16 +197,17 @@ steps:
- tests/v1
commands:
# split the test to avoid interference
- VLLM_USE_V1=1 pytest -v -s v1/core
- VLLM_USE_V1=1 pytest -v -s v1/engine
- VLLM_USE_V1=1 pytest -v -s v1/sample
- VLLM_USE_V1=1 pytest -v -s v1/worker
- VLLM_USE_V1=1 pytest -v -s v1/structured_output
- VLLM_USE_V1=1 pytest -v -s v1/test_stats.py
- VLLM_USE_V1=1 pytest -v -s v1/test_utils.py
- pytest -v -s v1/core
- pytest -v -s v1/engine
- pytest -v -s v1/sample
- pytest -v -s v1/worker
- pytest -v -s v1/structured_output
- pytest -v -s v1/test_stats.py
- pytest -v -s v1/test_utils.py
- pytest -v -s v1/test_oracle.py
# TODO: accuracy does not match, whether setting
# VLLM_USE_FLASHINFER_SAMPLER or not on H100.
- VLLM_USE_V1=1 pytest -v -s v1/e2e
- pytest -v -s v1/e2e
# Integration test for streaming correctness (requires special branch).
- pip install -U git+https://github.com/robertgshaw2-neuralmagic/lm-evaluation-harness.git@streaming-api
- pytest -v -s entrypoints/openai/correctness/test_lmeval.py::test_lm_eval_accuracy_v1_engine
@@ -226,12 +227,12 @@ steps:
- python3 offline_inference/llm_engine_example.py
- python3 offline_inference/vision_language.py
- python3 offline_inference/vision_language_multi_image.py
- python3 other/tensorize_vllm_model.py --model facebook/opt-125m serialize --serialized-directory /tmp/ --suffix v1 && python3 other/tensorize_vllm_model.py --model facebook/opt-125m deserialize --path-to-tensors /tmp/vllm/facebook/opt-125m/v1/model.tensors
- VLLM_USE_V1=0 python3 other/tensorize_vllm_model.py --model facebook/opt-125m serialize --serialized-directory /tmp/ --suffix v1 && python3 other/tensorize_vllm_model.py --model facebook/opt-125m deserialize --path-to-tensors /tmp/vllm/facebook/opt-125m/v1/model.tensors
- python3 offline_inference/encoder_decoder.py
- python3 offline_inference/basic/classify.py
- python3 offline_inference/basic/embed.py
- python3 offline_inference/basic/score.py
- python3 offline_inference/profiling.py --model facebook/opt-125m run_num_steps --num-steps 2
- VLLM_USE_V1=0 python3 offline_inference/profiling.py --model facebook/opt-125m run_num_steps --num-steps 2
- label: Prefix Caching Test # 9min
mirror_hardwares: [amd]
@@ -375,7 +376,8 @@ steps:
commands:
- pytest -v -s models/test_transformers.py
- pytest -v -s models/test_registry.py
- pytest -v -s models/test_initialization.py
# V1 Test: https://github.com/vllm-project/vllm/issues/14531
- VLLM_USE_V1=0 pytest -v -s models/test_initialization.py
- label: Language Models Test (Standard) # 32min
#mirror_hardwares: [amd]
@@ -518,8 +520,8 @@ steps:
# this test fails consistently.
# TODO: investigate and fix
# - pytest -v -s spec_decode/e2e/test_integration_dist_tp2.py
- CUDA_VISIBLE_DEVICES=0,1 pytest -v -s test_sharded_state_loader.py
- CUDA_VISIBLE_DEVICES=0,1 pytest -v -s kv_transfer/disagg_test.py
- VLLM_USE_V1=0 CUDA_VISIBLE_DEVICES=0,1 pytest -v -s test_sharded_state_loader.py
- VLLM_USE_V1=0 CUDA_VISIBLE_DEVICES=0,1 pytest -v -s kv_transfer/disagg_test.py
- label: Plugin Tests (2 GPUs) # 40min
working_dir: "/vllm-workspace/tests"