- 09 Feb, 2026 2 commits
- 06 Feb, 2026 1 commit
-
-
wanglong3 authored
feat: support moe sum when topk==9 bugfix: Fix mtp model load error when eable shared experts fusion.
-
- 08 Jan, 2026 5 commits
-
-
zhuwenwen authored
feat: Support enable rms quant and shared expert overlap at same time. See merge request dcutoolkit/deeplearing/vllm!352
-
wanglong3 authored
-
zhuwenwen authored
fix：更新 Marlin W16A16 MoE 中支持的断言消息 See merge request dcutoolkit/deeplearing/vllm!349
-
laibao authored
-
zhuwenwen authored
解决custom allreduce在dp情况下的其服务错误问题 See merge request dcutoolkit/deeplearing/vllm!348
-
- 07 Jan, 2026 4 commits
- 06 Jan, 2026 6 commits
- 05 Jan, 2026 5 commits
- 04 Jan, 2026 9 commits
-
-
zhuwenwen authored
V0.9.2 dev pdncclinit See merge request dcutoolkit/deeplearing/vllm!333
-
xiabo authored
-
xiabo authored
-
zhuwenwen authored
fix: 只有当kv block中不含有MTP的假数据时才会被cached，以修复cache_full_blocks同一个kv block保存两次的bug See merge request dcutoolkit/deeplearing/vllm!331
-
zhuwenwen authored
perf: 加速 v1 InputBatch.add_request 的 token_ids 拷贝 See merge request dcutoolkit/deeplearing/vllm!330
-
jujl1 authored
-
laibao authored
新增环境变量开关 VLLM_V1_FAST_TOKEN_ID_COPY（默认关闭）。开启后在 CachedRequestState 中缓存 prompt_token_ids 的 np.int32，并在 add_request 里用 np.copyto 写入 token_ids_cpu，避免长 prompt 场景反复 list->NumPy 转换开销（尤其是抢占/反复进出 batch 时）
-
zhuwenwen authored
perf: DS-量化模型融合qa和kva的gemm See merge request dcutoolkit/deeplearing/vllm!329
-
wujl5 authored
-
- 31 Dec, 2025 8 commits
-
-
gaoqiong authored
fix: chunk-prefill生成的token被错误返回 See merge request dcutoolkit/deeplearing/vllm!327
-
yangql authored -
jujl1 authored
-
yangql authored
[feat]解决deepep auto模式 mtp >1 时卡住 See merge request dcutoolkit/deeplearing/vllm!326
-
-
王敏 authored
-
gaoqiong authored
V0.9.2 dev main+mtp zero See merge request dcutoolkit/deeplearing/vllm!325
-
yangql authored
-
