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

[Bugfix] fix pcp qwen full graph FIA bug (#6037)

Browse Source 
### What this PR does / why we need it?
In the pcp full graph Qwen model scenario, the inconsistency between the
Q shape and actual q len of the FIA operator is fixed.

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

### How was this patch tested?

- vLLM version: v0.13.0
- vLLM main:
2c24bc6996

Signed-off-by: weiguihua2 <weiguihua2@huawei.com>
This commit is contained in:
weiguihua2
2026-01-21 08:49:05 +08:00
committed by GitHub
parent b6d55fc48e
commit b399117e89
1 changed files with 2 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
vllm_ascend/compilation/acl_graph.py
View File

@@ -440,11 +440,8 @@ def update_attn_dcp_pcp_params(update_stream, forward_context, runtime_shape):
pad_tensor = np.zeros(pad_length, dtype=actual_seq_lengths_kv.dtype) pad_tensor = np.zeros(pad_length, dtype=actual_seq_lengths_kv.dtype)
actual_seq_lengths_kv = np.concatenate([actual_seq_lengths_kv, pad_tensor]) actual_seq_lengths_kv = np.concatenate([actual_seq_lengths_kv, pad_tensor])
actual_seq_lengths_q = attn_metadata.actual_seq_lengths_q[: attn_metadata.num_decode_tokens] actual_seq_lengths_q = attn_metadata.actual_seq_lengths_q
if runtime_shape - len(actual_seq_lengths_q):
actual_seq_lengths_q = actual_seq_lengths_q + [actual_seq_lengths_q[-1]] * (
runtime_shape - len(actual_seq_lengths_q)
)
if dcp_size > 1: if dcp_size > 1:
num_heads = num_heads * dcp_size num_heads = num_heads * dcp_size