Lehua Ding e184c9c510 [perf] Use CPU tensor to reduce GPU->CPU sync (#25884) ...

Signed-off-by: Lehua Ding <lehuading@tencent.com>

2025-09-30 19:51:16 +08:00

..

__init__.py

[V1] Implement vLLM V1 [1/N] (#9289)

2024-10-22 01:24:07 -07:00

block_table.py

[Core] Use CpuGpuBuffer for block table tensors (#24795)

2025-09-16 19:18:06 -07:00

cpu_model_runner.py

[Core] Modify the initialization parameters of the lora manager (#25249)

2025-09-19 18:01:28 +00:00

cpu_worker.py

[CPU] update torch 2.8 and fix missing fields in TorchSDPAMetadata (#25652)

2025-09-25 13:29:11 +00:00

gpu_input_batch.py

[Bugfix] fix apply_temperature to avoid nan in probs (#24734)

2025-09-25 05:32:21 +00:00

gpu_model_runner.py

[perf] Use CPU tensor to reduce GPU->CPU sync (#25884)

2025-09-30 19:51:16 +08:00

gpu_ubatch_wrapper.py

Reduce the Cuda Graph memory footprint when running with DBO (#25779)

2025-09-26 22:29:56 +00:00

gpu_worker.py

[Logging] Remove TORCH_NCCL_AVOID_RECORD_STREAMS to squash a warning (#25532)

2025-09-25 15:16:06 +00:00

kv_connector_model_runner_mixin.py

[Bugfix] Fix missing clear_connector_metadata (#25397)

2025-09-23 08:10:59 +08:00

lora_model_runner_mixin.py

[Core] Modify the initialization parameters of the lora manager (#25249)

2025-09-19 18:01:28 +00:00

tpu_input_batch.py

[CORE] Prompt Embeddings Support for v1 Engine (#24278)

2025-09-19 08:03:09 +08:00

tpu_model_runner.py

[New Model] DeepSeek-V3.2 (Rebased to Main) (#25896)

2025-09-30 17:14:41 +08:00

tpu_worker.py

[TPU][Bugfix] fix the missing apply_model in tpu worker (#25526)

2025-09-24 05:18:08 +00:00

ubatch_splitting.py

[BugFix] Fix using dbo_decode_token_threshold always (and ignoring dbo_prefill_token_threshold) (#25622)

2025-09-26 16:22:49 +00:00

ubatch_utils.py

[Core/DBO][2/N] Dual-Batch Overlap add DeepEP High Throughput support and Prefill support (#24845)

2025-09-23 16:02:10 +00:00

ubatching.py

[Core/DBO][2/N] Dual-Batch Overlap add DeepEP High Throughput support and Prefill support (#24845)

2025-09-23 16:02:10 +00:00

utils.py

[Model] Add LongCat-Flash (#23991)

2025-09-24 21:53:40 -07:00

worker_base.py

[V0 Deprecation] Remove vllm.worker and update according imports (#25901)

2025-09-29 23:26:11 +00:00

xpu_model_runner.py

[XPU] Fix xpu model runner call torch.cuda APIs (#25011)

2025-09-17 06:45:25 +00:00

xpu_worker.py

[Doc]: fix typos in Python comments (#24115)

2025-09-02 21:14:07 -07:00