Andreas Karatzas 66f927f205 [Bugfix] Fix pooling non-determinism from pinned prompt_lens aliasing (#37775) ...

Signed-off-by: Andreas Karatzas <akaratza@amd.com>

2026-03-22 03:22:24 +00:00

..

__init__.py

[V1] Adding min tokens/repetition/presence/frequence penalties to V1 sampler (#10681)

2024-12-26 19:02:58 +09:00

test_gpu_input_batch.py

[Bugfix] Fix pooling non-determinism from pinned prompt_lens aliasing (#37775)

2026-03-22 03:22:24 +00:00

test_gpu_model_runner.py

[NIXL][1/N] Refactor kernel_block_size detection (#35752)

2026-03-11 01:11:23 -07:00

test_gpu_profiler.py

Support custom URI schemes and trace handlers for profiler (#32393)

2026-01-22 09:45:40 -08:00

test_late_interaction_runner.py

[Perf] Optimize compute maxsim using batched version, 3.2% E2E throughput improvement (#36710)

2026-03-12 08:37:01 +08:00

test_mamba_utils.py

[Hybrid] calling get_mamba_groups() once at MambaCopyBuffers.create() (#37318)

2026-03-21 09:29:43 +00:00

test_utils.py

[5/N][Attention] Finish eliminating vllm/attention folder (#32064)

2026-01-27 10:02:51 -05:00

test_worker_memory_snapshot.py

[Hardware] Replace torch.cuda.device_count/current_device/set_device API (#36145)

2026-03-12 07:57:47 -07:00