vllm/attention/ops/chunked_prefill_paged_decode.py · 1a19e9cd87b6bb091fc9c530e6446601db52b58e · OpenDAS / vllm_cscc

[Bugfix][ROCm]Fix Qwen3-Next-80B-A3B-Thinking inference and optimize... · 1a19e9cd

vllmellm authored Jan 09, 2026


[Bugfix][ROCm]Fix Qwen3-Next-80B-A3B-Thinking inference and optimize non-standard block size (544) support under rocm_atten (#31380)
Signed-off-by: vllmellm <vllm.ellm@embeddedllm.com>

1a19e9cd

chunked_prefill_paged_decode.py 15.3 KB

Replace chunked_prefill_paged_decode.py