biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Bugfix][CI] fix typos (#34934)

Browse Source 
Signed-off-by: 1195343015 <1195343015@qq.com>
Signed-off-by: Jiayi Yan <66017932+1195343015@users.noreply.github.com>
Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
This commit is contained in:
Jiayi Yan
2026-03-06 01:05:46 +08:00
committed by GitHub
parent 8c760b6ab6
commit 6a895197fa
98 changed files with 227 additions and 366 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
tests/models/multimodal/generation/vlm_utils/model_utils.py
View File

@@ -719,7 +719,7 @@ def isaac_patch_hf_runner(hf_model: HfRunner) -> HfRunner:
# Convert to tuple or None
all_hidden_states = tuple(hidden_states_list) if output_hidden_states else None
# Include hiden_states for compatibility with hidden_states_to_seq_logprobs()
# Include hidden_states for compatibility with hidden_states_to_seq_logprobs()
return BaseModelOutputWithPast(
last_hidden_state=hidden_states,
past_key_values=past_key_values,
@@ -1226,7 +1226,7 @@ def voxtral_patch_hf_runner(hf_model: "HfRunner") -> "HfRunner":
dicts (accepting ``url``, ``path``, or ``base64`` audio) rather than
the standard ``processor(text=, audio=, sampling_rate=)`` interface.
2. HfRunner.get_inputs cannot handle multi-audio per prompt because it
mis-unpacks ``[(arr1, sr1), (arr2, sr2)]`` via a ``len == 2`` check.
incorrectly unpacks ``[(arr1, sr1), (arr2, sr2)]`` via a ``len == 2`` check.
We override ``get_inputs`` to build conversation dicts and call
``apply_chat_template`` directly, bypassing both issues. We also wrap