Actions: vllm-project/vllm-ascend
Actions
1,169 workflow runs
1,169 workflow runs
FULL_DECODE_ONLY
mode for GQA/MHA models
ascend test / full
#740:
Pull request #2128
synchronize
by
yiz-liu
deepseek mtp
, enable_shared_expert_dp
and use_cached_kv_cache_bytes
ascend test / full
#726:
Pull request #3074
labeled
by
linfeng-yuan
ProTip!
You can narrow down the results and go further in time using created:<2025-09-22 or the other filters available.