EngineX/xc-llm-ascend
3
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

[1/N][CI] Move linting system to pre-commits hooks (#1256)

Browse Source 
### What this PR does / why we need it?

Follow vllm-project/vllm lint way:
https://github.com/vllm-project/vllm/blob/main/.pre-commit-config.yaml

Enable pre-commit to avoid some low level error  AMAP.

This pr is one step of #1241, The purpose is make linting system more
clear and convenient, on this step, Mainly did the following things:
yapf, actionlint, ruff, typos, isort, mypy, png-lint, signoff-commit,
enforce-import-regex-instead-of-re.

TODO: 
- clang-format(check for csrc with google style)
need clean code, disable for now 
- pymarkdown
need clean code, disable for now 
- shellcheck
need clean code, disable for now 

### Does this PR introduce _any_ user-facing change?

Only developer UX change:

https://vllm-ascend--1256.org.readthedocs.build/en/1256/developer_guide/contributing.html#run-lint-locally

```
pip install -r requirements-lint.txt && pre-commit install
bash format.sh
```

### How was this patch tested?

CI passed with new added/existing test.

Co-authored-by: Yikun [yikunkero@gmail.com](mailto:yikunkero@gmail.com)
Co-authored-by: wangli
[wangli858794774@gmail.com](mailto:wangli858794774@gmail.com)
- vLLM version: v0.9.1
- vLLM main:
5358cce5ff

---------

Signed-off-by: wangli <wangli858794774@gmail.com>
This commit is contained in:
Li Wang
2025-07-10 14:17:15 +08:00
committed by GitHub
parent 643e6f5486
commit c7446438a9
28 changed files with 753 additions and 667 deletions
Download Patch File Download Diff File Expand all files Collapse all files

23
vllm_ascend/ops/fused_moe.py
View File

@@ -79,8 +79,9 @@ def process_topk_ids(topk_ids: torch.Tensor, expert_num: int, ep_size: int,
experts_per_ep_rank_val).to(original_dtype)
indices_arange = torch.arange(topk_ids.shape[0], device=device)
is_new_segment = torch.cat((torch.tensor([True], device=device),
assigned_ep_rank[1:] != assigned_ep_rank[:-1]))
is_new_segment = torch.cat(
(torch.tensor([True], device=device), assigned_ep_rank[1:]
!= assigned_ep_rank[:-1]))
temp_start_markers = torch.full_like(indices_arange,
-1,
dtype=indices_arange.dtype)
@@ -469,13 +470,13 @@ def fused_experts_with_all2all_buffer(
expert_idx_buffer_scatter.shape,
dtype=expert_idx_buffer_scatter.dtype,
device=expert_idx_buffer_scatter.device)
non_pad_len = torch.sum(
(expert_idx_buffer_scatter != global_num_experts).to(torch.int32))
hidden_states_pad_idx[
expert_idx_buffer_scatter != global_num_experts] = torch.arange(
non_pad_len,
dtype=expert_idx_buffer_scatter.dtype,
device=hidden_states.device)
non_pad_len = torch.sum((expert_idx_buffer_scatter
!= global_num_experts).to(torch.int32))
hidden_states_pad_idx[expert_idx_buffer_scatter !=
global_num_experts] = torch.arange(
non_pad_len,
dtype=expert_idx_buffer_scatter.dtype,
device=hidden_states.device)
hidden_states_buffer_scatter = hidden_states[hidden_states_pad_idx]
expert_idx_buffer_gather = torch.empty_like(
@@ -528,8 +529,8 @@ def fused_experts_with_all2all_buffer(
dist.all_to_all_single(hidden_states_gatter,
hidden_states_scatter,
group=ep_group.device_group)
hidden_states_gatter = hidden_states_gatter[
expert_idx_buffer_scatter != global_num_experts]
hidden_states_gatter = hidden_states_gatter[expert_idx_buffer_scatter !=
global_num_experts]
if hidden_states_gatter.shape[0] != row_idx_len:
hidden_states = torch.zeros((row_idx_len, hidden_states.shape[1]),
dtype=hidden_states.dtype,

4
vllm_ascend/ops/vocab_parallel_embedding.py
View File

@@ -30,8 +30,8 @@ def get_masked_input_and_mask(
added_vocab_end_index: int) -> Tuple[torch.Tensor, torch.Tensor]:
# torch.compile will fuse all of the pointwise ops below
# into a single kernel, making it very fast
org_vocab_mask = (input_ >= org_vocab_start_index) & (input_ <
org_vocab_end_index)
org_vocab_mask = (input_ >= org_vocab_start_index) & (
input_ < org_vocab_end_index)
added_vocab_mask = (input_ >= added_vocab_start_index) & (
input_ < added_vocab_end_index)
added_offset = added_vocab_start_index - (