weijinqian0 dbe4c338f2 [Refactor] cache cos/sin in mla & remove parameter model in builder. (#5277) ...

RFC: https://github.com/vllm-project/vllm-ascend/issues/4629

1. Cache cos/sin in mla
2. AttentionBuilder inherits from the original class of vllm.



version: release/v0.13.0
- vLLM main:
ad32e3e19c

---------

Signed-off-by: weijinqian_v1 <weijinqian@huawei.com>
Co-authored-by: weijinqian_v1 <weijinqian@huawei.com>

2025-12-28 10:35:07 +08:00

..

fused_moe

Revert "[feat] enable hierarchical mc2 ops on A2 by default (#5300)" (#5434)

2025-12-27 17:06:58 +08:00

triton

rollback causal_conv1d_fn to torch ops & update qwen3Next doc (#5391)

2025-12-26 19:57:38 +08:00

__init__.py

[Fusion] [Graph] Add qknorm rope fusion operator (#4711)

2025-12-17 08:53:44 +08:00

activation.py

[refact] unified soc_version code (#4359)

2025-11-26 14:28:55 +08:00

expert_load_balancer.py

eplb redundant expert bugfix (#4291)

2025-11-21 14:24:35 +08:00

layernorm.py

[Graph][Fusion]Add new pattern for AddRmsnormQuant with SP. (#5077)

2025-12-18 20:25:44 +08:00

linear_op.py

Remove VLLM_ASCEND_ENABLE_DENSE_OPTIMIZE (#5272)

2025-12-25 11:09:56 +08:00

linear.py

[bugfix] fix Error 'ValueError: Duplicate layer name' (#5280)

2025-12-25 10:43:24 +08:00

mla.py

upgrade vLLM to main (#4608)

2025-12-02 22:10:52 +08:00

mm_encoder_attention.py

[Bugfix] Correctly handle the output shape in multimodal attention (#5443)

2025-12-27 18:42:46 +08:00

register_custom_ops.py

[refactor] Remove unnecessary attributes from set_ascend_forward_context (#5204)

2025-12-23 08:49:52 +08:00

rotary_embedding.py

[Refactor] cache cos/sin in mla & remove parameter model in builder. (#5277)

2025-12-28 10:35:07 +08:00

shared_weight_layer.py

[Feat] Flashcomm2 use o_shared linear (#4188)

2025-12-11 12:43:04 +08:00

vocab_parallel_embedding.py

[Feat] Add custom Embedding tensor model parallel (#2616)

2025-12-12 14:41:20 +08:00

weight_prefetch.py

Update torch-npu version to 2.7.1 (#3896)

2025-10-31 17:16:31 +08:00