karylab_agents/vllm

Fork 0

forked from Karylab-cklius/vllm

Files

T

History

8521175db9 fix: migrate gpt_oss_20b moe_backend from env var to model kwargs

- Add `flashinfer_trtllm_afp8` and `flashinfer_cutlass_afp8` to MoEBackend
  Literal in vllm/config/kernel.py so they are valid kernel_config values
- Attach `moe_backend="flashinfer_trtllm_afp8"` to the gpt_oss_20b model
  object in models.py (conditionally on Blackwell) instead of setting the
  env var in the test runner
- Remove the TODO env var block from conftest.py

Co-authored-by: GitHub Copilot

Agent-Logs-Url: https://github.com/vllm-project/vllm/sessions/8abf6884-3fc2-45fb-b5c8-1ba82b14c841

Co-authored-by: ProExpertProg <11367180+ProExpertProg@users.noreply.github.com>

2026-04-29 18:15:03 +00:00

correctness_e2e

[torch.compile]: Disable Sequence Parallelism (SP) for piecewise compilation (#38373 )

2026-04-26 17:44:42 +00:00

fullgraph

full cudagraph for flex-attn (#36298 )

2026-04-02 21:15:01 -07:00

fusions_e2e

fix: migrate gpt_oss_20b moe_backend from env var to model kwargs

2026-04-29 18:15:03 +00:00

h100

[Core] Account for num_gpu_blocks_override in max_model_len checks (#41069 )

2026-04-28 15:44:09 -07:00

passes

[torch.compile]: Disable Sequence Parallelism (SP) for piecewise compilation (#38373 )

2026-04-26 17:44:42 +00:00

__init__.py

[torch.compile] register allreduce operations as custom ops (#8526 )

2024-09-16 22:57:57 -07:00

backend.py

[vLLM IR] 1/N Implement IR skeleton and rms_norm op (#33825 )

2026-03-31 22:15:05 -04:00

conftest.py

[torch.compile] Sequence Parallelism threshold compile ranges (#28672 )

2026-02-26 05:00:12 +00:00

README.md

[CI] Reorganize compile tests so new tests are automatically included in CI (#28625 )

2025-11-19 06:13:50 -08:00

silly_attention.py

[Chore] Clean up pytorch helper functions in vllm.utils (#26908 )

2025-10-18 09:48:22 -07:00

test_aot_compile.py

[BugFix] fix VLLM_USE_STANDALONE_COMPILE=0 (#38015 )

2026-03-24 19:08:26 +00:00

test_compile_ranges.py

[torch.compile] Use FakeTensors instead of real GPU tensors for single-size compilation (#36093 )

2026-03-11 16:07:09 +00:00

test_config.py

[Bugfix] Fix max_num_batched_token not captured in cuda graph (#40734 )

2026-04-28 21:33:06 -07:00

test_decorator.py

[Bugfix][CI] fix typos (#34934 )

2026-03-05 17:05:46 +00:00

test_dynamic_shapes_compilation.py

[Test] Fix test_dynamic_shapes_compilation for torch 2.12 (#40743 )

2026-04-27 17:51:15 -07:00

test_graph_partition.py

[Test] Refactor hard coded device string in test files under compile/quantization/models/model_executor folders (#38901 )

2026-04-15 11:02:35 +08:00

test_rotary_embedding_compile.py

[Test] Refactor hard coded device string in test files under compile/quantization/models/model_executor folders (#38901 )

2026-04-15 11:02:35 +08:00

test_sequence_parallelism_threshold.py

[torch.compile] Sequence Parallelism threshold compile ranges (#28672 )

2026-02-26 05:00:12 +00:00

test_structured_logging.py

[Test] Refactor hard coded device string in test files under compile/quantization/models/model_executor folders (#38901 )

2026-04-15 11:02:35 +08:00

test_wrapper.py

[Bugfix][CI] fix typos (#34934 )

2026-03-05 17:05:46 +00:00

README.md

compile test folder structure

compile/test_*.py : various unit tests meant for testing particular code path/features. Future tests are most likely added here. New test files added here will be included in CI automatically
compile/fullgraph/ : full model tests, including all tests previously in compile/piecewise. These tests do not target particular features. New test files added here will be included in CI automatically
compile/distributed/ : tests that require multiple GPUs. New test files added here will NOT be included in CI automatically as these tests generally need to be manually configured to run in runners with particular number/type of GPUs.