Wang Yixuan 5d8ec28009 [2/N][refactor] split torchair from fused_moe (#2503) ...

### What this PR does / why we need it?
After moved torchair related fused_moe section into torchair_fused_moe,
split the torchair from the origin fused_moe

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

### How was this patch tested?
vLLM version: main
vLLM main:
ab9f2cfd19


- vLLM version: v0.10.1.1
- vLLM main:
2a97ffc33d

Signed-off-by: hust17yixuan <303660421@qq.com>

2025-08-26 14:12:43 +08:00

..

layers

refactor select_experts of moe module (#2150)

2025-08-14 11:50:53 +08:00

moe_dispatcher

[main] refactor alltoallv in fused_moe (#2487)

2025-08-23 20:38:17 +08:00

__init__.py

[CustomOp] Register RotaryEmbedding instead of overwrite forward (#2385)

2025-08-25 09:32:35 +08:00

activation.py

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

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

attention.py

Disaggregate prefill for kv cache register style (#950)

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

comm_utils.py

[main][Feature]Moe alltoallv communication optimization for unquantized RL training sence (#2088)

2025-08-02 09:49:10 +08:00

common_fused_moe.py

refactor select_experts of moe module (#2150)

2025-08-14 11:50:53 +08:00

expert_load_balancer.py

Add static EPLB (#1116)

2025-06-09 19:28:11 +08:00

fused_moe.py

[2/N][refactor] split torchair from fused_moe (#2503)

2025-08-26 14:12:43 +08:00

layernorm.py

[CustomOp] Register RMSNorm instead of overwrite forward_oot (#2284)

2025-08-14 17:18:30 +08:00

linear.py

add mlp tp optimze (#2120)

2025-08-21 09:22:07 +08:00

rotary_embedding.py

[CustomOp] Register RotaryEmbedding instead of overwrite forward (#2385)

2025-08-25 09:32:35 +08:00

sequence_parallel.py

【main】SP For Qwen3 MoE (#2209)

2025-08-07 09:15:49 +08:00

vocab_parallel_embedding.py

[V1] MTP supports torchair (#2145)

2025-08-06 19:37:43 +08:00