Clorist33 2b5b309133 [Bugfix]Fix precision issues in moe_mlp (vllm-ascend v0.11.0-dev) (#5023) ...

### What this PR does / why we need it?
Use group_list[0] to replace group_diff[0] in function
"cumsum_group_list" (moe_mlp.py).
The purpose is to modify it to the correct logic of converting cumsum to
count.

### Does this PR introduce _any_ user-facing change?
No

Signed-off-by: tanqingshan (A)  <50050625@china.huawei.com>
Co-authored-by: tanqingshan (A) <50050625@china.huawei.com>

2025-12-16 08:40:03 +08:00

..

attention

[Cherry-pick]bmm_transpose to v011dev (#3995)

2025-12-08 19:22:14 +08:00

compilation

[v0.11.0-dev][bugfix] Add branch for stream up-lifting in update_attn_params (#4437)

2025-12-08 08:54:46 +08:00

core

[BugFix][Cherry-pick] Cherry-pick PR 3675 to v0.11.0-dev (#3732)

2025-10-25 09:41:51 +08:00

device_allocator

[Misc]Clean up useless import from vllm (#2049)

2025-07-28 16:01:59 +08:00

distributed

[0.11.0][Bugfix] Remove the ZMQ communication setup on the D node (#4916)

2025-12-12 14:37:49 +08:00

eplb

BugFix: Resolve shape mismatch in eplb update and calculation issues in quant_apply_mlp (#4777)

2025-12-09 15:46:58 +08:00

lora

[Bugfix][LoRA] Fix forward error and shape mismatch when using LoRA (#3153)

2025-09-28 17:30:50 +08:00

models

[Bugfix] fix qwen3-vl-moe shape ERROR during the _prepare_inputs phase under high concurrency. (#4658)

2025-12-08 19:30:16 +08:00

multistream

[Quickfix] update CachedRequestState as NewRequestData changed (#2367)

2025-08-15 07:35:27 +08:00

ops

[Bugfix]Fix precision issues in moe_mlp (vllm-ascend v0.11.0-dev) (#5023)

2025-12-16 08:40:03 +08:00

patch

[Bugfix][SHM] Use writer lock by default and remove redundant env (#4117)

2025-12-01 22:27:01 +08:00

quantization

fix nz for quantization (#4943)

2025-12-12 14:54:41 +08:00

sample

[0.11.0][ops] npu_top_k_top_p supports k and p only (#4153)

2025-12-09 15:45:40 +08:00

spec_decode

[Bugfix] Fix the Eagle3 inference failure issue. (#4721)

2025-12-12 14:52:29 +08:00

torchair

【0.11.0-dev】optimization of kimi-k2 in cann8.3 (#4555)

2025-12-09 08:49:15 +08:00

worker

[Bugfix] fix qwen3-vl-moe shape ERROR during the _prepare_inputs phase under high concurrency. (#4658)

2025-12-08 19:30:16 +08:00

__init__.py

[Refactor] Adapt deepseek-v3.2 to vllm 0.11.0 (#3432)

2025-10-15 17:48:58 +08:00

ascend_config.py

[main] support cpu binding (#3546)

2025-10-21 09:17:03 +08:00

ascend_forward_context.py

Revert "[refactor]support gatingtopk operator generalization (#4356)" (#4873)

2025-12-10 15:45:20 +08:00

cpu_binding.py

[main] support cpu binding (#3546)

2025-10-21 09:17:03 +08:00

envs.py

[Cherry-pick] [0.11.0] pd proxy support ipv6 and fix proxy (#4242)

2025-11-18 16:33:00 +08:00

meta_registration.py

Fix the bugs about operator registration by PyTorch Dispatcher (#2786)

2025-09-13 11:58:52 +08:00

platform.py

[MM][Bugfix] Add error log for VL models when enabling FLASHCOMM (#4222)

2025-11-21 15:04:35 +08:00

utils.py

fix nz for quantization (#4943)

2025-12-12 14:54:41 +08:00