Robert Shaw c777df79f7 [BugFix] Fix Memory Leak (#17567) ...

Signed-off-by: rshaw@neuralmagic.com <robertgshaw2@gmail.com>

2025-05-02 01:07:03 -07:00

..

attention

[Attention] MLA move o_proj q_proj into cuda-graph region (#17484)

2025-05-02 03:16:26 +00:00

core

[BugFix] Fix Memory Leak (#17567)

2025-05-02 01:07:03 -07:00

engine

[V1][Metrics] add support for kv event publishing (#16750)

2025-04-30 07:44:45 -07:00

executor

[BugFix] Remove default multiproc executor collective_rpc timeout (#17000)

2025-04-22 23:27:14 +00:00

metrics

[V1][Bugfix]: vllm v1 verison metric num_gpu_blocks is None (#15755)

2025-04-30 18:20:39 +08:00

sample

[V1][Spec Decode] Always use argmax for sampling draft tokens (#16899)

2025-04-23 14:57:43 -07:00

spec_decode

[V1][Spec Decode] Apply torch.compile & cudagraph to EAGLE3 (#17504)

2025-05-01 16:19:30 -07:00

stats

Update deprecated Python 3.8 typing (#13971)

2025-03-02 17:34:51 -08:00

structured_output

[V1][Feature] Enable Speculative Decoding with Structured Outputs (#14702)

2025-04-30 00:02:10 +00:00

worker

[V1][Feature] Enable Speculative Decoding with Structured Outputs (#14702)

2025-04-30 00:02:10 +00:00

__init__.py

[V1] AsyncLLM Implementation (#9826)

2024-11-11 23:05:38 +00:00

kv_cache_interface.py

[V1] Implement sliding window attention in kv_cache_manager (#14097)

2025-04-01 00:33:17 -07:00

outputs.py

[V1] Aggregate chunked prompt logprobs in model runner (#14875)

2025-03-24 12:27:57 -04:00

request.py

[Core] Prevent side-channel attacks via cache salting (#17045)

2025-04-30 20:27:21 +08:00

serial_utils.py

[V1] Allow turning off pickle fallback in vllm.v1.serial_utils (#17427)

2025-04-30 16:10:54 +00:00

utils.py

[V1] Move usage stats to worker and start logging TPU hardware (#16211)

2025-04-25 14:06:01 -06:00