huangxialu 9c9a7cd90b [main] adapt usage of npu_moe_gating_top_k_softmax and remove envs.SELECT_GATING_TOPK_SOTFMAX_EXPERTS (#2112) ...

backport of v0.9.1-dev:
https://github.com/vllm-project/vllm-ascend/pull/1902

origin main npu_moe_gating_top_k_softmax:
https://github.com/vllm-project/vllm-ascend/pull/1355

- vLLM version: v0.10.0
- vLLM main:
055bd3978e

Signed-off-by: huangxialu <huangxialu1@huawei.com>

2025-07-31 21:05:56 +08:00

..

attention

[Perf] Avoid performing index selection of sin/cos cache every layer (#1890)

2025-07-29 18:06:45 +08:00

core

[Test] Add uts for files in /core (#1957)

2025-07-25 09:48:19 +08:00

device_allocator

add ut for device allocator/camem and mutistream/layers (#2037)

2025-07-31 19:17:27 +08:00

distributed/device_communicators

[Misc]Remove PD v0 code (#2047)

2025-07-28 19:09:22 +08:00

fake_weight

[CI] Add unit test framework (#1201)

2025-06-16 18:32:28 +08:00

kv_connector

Disaggregate prefill for kv cache register style (#950)

2025-07-26 17:15:47 +08:00

models

ut:add ut for qwen2_5_vl (#2143)

2025-07-31 20:46:17 +08:00

multistream

add ut for device allocator/camem and mutistream/layers (#2037)

2025-07-31 19:17:27 +08:00

ops

[main] adapt usage of npu_moe_gating_top_k_softmax and remove envs.SELECT_GATING_TOPK_SOTFMAX_EXPERTS (#2112)

2025-07-31 21:05:56 +08:00

patch/worker/patch_common

[Refactor]Refactor sampler (#2050)

2025-07-30 08:47:22 +08:00

quantization

[main][Feature] Support Qwen3 W4A8 quantization (#2060)

2025-07-30 14:57:14 +08:00

sample

[Refactor]Refactor sampler (#2050)

2025-07-30 08:47:22 +08:00

torchair

[2/4][Refactor] Refactor torchair utils (#1892)

2025-07-21 19:43:30 +08:00

worker

[FOLLOWUP] Use base test to avoid patch everwhere (#1634)

2025-07-22 09:03:40 +08:00

__init__.py

[2/4][Refactor] Refactor torchair utils (#1892)

2025-07-21 19:43:30 +08:00

base.py

[Feature]: implement the fusion of allreduce and matmul in prefill phase when tp is enabled (#1926)

2025-07-28 15:13:37 +08:00

conftest.py

[1/N][CustomOp] Register activation customop instead of overwrite forward_oot (#1841)

2025-07-18 23:07:14 +08:00

test_ascend_config.py

[Dist][EP] Remove ETP/EP maintained in vllm-ascend (#1681)

2025-07-21 09:08:04 +08:00

test_platform.py

[1/4][Refactor] Refactor torchair worker (#1885)

2025-07-21 11:50:46 +08:00

test_utils.py

[Bugfix] Fix num_hidden_layers when Qwen2-Audio 7B (#1803)

2025-07-26 20:13:00 +08:00