Self Checks
Dify version
1.13.3
Plugin version
0.1.41
Cloud or Self Hosted
Self Hosted (Source)
Steps to reproduce
- Configure a model (e.g.,
multimodal-embedding-v1) that has a valid YAML configuration file in its directory.
- The YAML file contains valid definitions, specifically including
'features': ['vision'].
- Trigger the logic that calls the
_is_vision_model method to check if the model supports vision features.
- Observe the console output and the returned value.
✔️ Error log
The method always returns False for vision models, completely ignoring the actual content of the YAML file.
This is caused by a logic flaw in the _is_vision_model method. When the YAML file is successfully parsed and the if condition for 'vision' is met, it correctly sets vision_models[model] = True. However, the code execution continues out of the try...except block, where an unconditional vision_models[model] = False overwrites the previously assigned True value.
As a result, any model with a valid YAML configuration is falsely flagged as not having vision capabilities, which breaks downstream multimodal features.
Self Checks
Dify version
1.13.3
Plugin version
0.1.41
Cloud or Self Hosted
Self Hosted (Source)
Steps to reproduce
multimodal-embedding-v1) that has a valid YAML configuration file in its directory.'features': ['vision']._is_vision_modelmethod to check if the model supports vision features.✔️ Error log
The method always returns
Falsefor vision models, completely ignoring the actual content of the YAML file.This is caused by a logic flaw in the
_is_vision_modelmethod. When the YAML file is successfully parsed and theifcondition for'vision'is met, it correctly setsvision_models[model] = True. However, the code execution continues out of thetry...exceptblock, where an unconditionalvision_models[model] = Falseoverwrites the previously assignedTruevalue.As a result, any model with a valid YAML configuration is falsely flagged as not having vision capabilities, which breaks downstream multimodal features.