[Bugfix] Fix more multimodal tests for transformers V5 (#34334)

Signed-off-by: raushan <raushan@huggingface.co>
This commit is contained in:
Raushan Turganbay
2026-02-11 22:02:05 +01:00
committed by GitHub
parent 5458eb835d
commit 527ca32197
5 changed files with 18 additions and 11 deletions
+3 -3
View File
@@ -810,9 +810,9 @@ class GlmAsrMultiModalProcessor(BaseMultiModalProcessor["GlmAsrProcessingInfo"])
# Postprocess: rename mask and add chunk counts
# Handle different key names from different transformers versions
if "input_feature_mask" in outputs:
outputs["feature_attention_mask"] = outputs.pop("input_feature_mask")
elif "feature_attention_mask" not in outputs and "input_features" in outputs:
if "input_features_mask" in outputs:
outputs["feature_attention_mask"] = outputs.pop("input_features_mask")
elif "input_features_mask" not in outputs and "input_features" in outputs:
# If no mask is provided, create one from input_features
input_features = outputs["input_features"]
if isinstance(input_features, torch.Tensor):