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

[MM][Bugfix] Update hf_config to hf_text_config (#5319)

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

Following https://github.com/vllm-project/vllm-ascend/pull/5205, update
`hf_config` to `hf_text_config`.

Find more details at
https://github.com/vllm-project/vllm-ascend/pull/5205#issuecomment-3675417534
and
https://github.com/vllm-project/vllm-ascend/pull/5205#issuecomment-3677920872.

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

### How was this patch tested?

- vLLM version: release/v0.13.0
- vLLM main:
5fbfa8d9ef

Signed-off-by: shen-shanshan <467638484@qq.com>
This commit is contained in:
Shanshan Shen
2026-01-06 16:41:39 +08:00
committed by GitHub
parent 293b2275df
commit b94d589769
23 changed files with 44 additions and 43 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
vllm_ascend/ascend_config.py
View File

@@ -96,7 +96,7 @@ class AscendConfig:
try:
# only support Qwen model now
# TODO: use a more robust method to get kv_head_num
num_kv_head = vllm_config.model_config.hf_config.num_key_value_heads
num_kv_head = vllm_config.model_config.hf_text_config.num_key_value_heads
self.num_head_replica = prefill_tp_size // num_kv_head if prefill_tp_size >= num_kv_head else 1
prefill_tp_size = min(prefill_tp_size, num_kv_head)
decode_tp_size = min(decode_tp_size, num_kv_head)
@@ -126,7 +126,7 @@ class AscendConfig:
self.enable_kv_nz = additional_config.get("enable_kv_nz", False)
if self.enable_kv_nz:
use_sparse = hasattr(vllm_config.model_config.hf_config,
use_sparse = hasattr(vllm_config.model_config.hf_text_config,
"index_topk")
if not vllm_config.model_config.is_deepseek_mla or use_sparse:
raise RuntimeError(