transformers - 💡(How to fix) Fix Qwen3_5 goes into infinite loop for a specific image [2 comments, 2 participants]

TL;DR

The issue can likely be resolved by preprocessing the specific image "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG" to match the expected dimensions handled by the qwen3_5 model.

Guidance

• Investigate the image dimensions of "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG" and compare them with the dimensions of images that work fine with the model.
• Preprocess the problematic image to resize it to the expected dimensions before passing it to the model for inference.
• Verify that the model handles images of different dimensions correctly by testing it with images of various sizes.
• Check the model's documentation or source code to see if there are any specific requirements or limitations for image dimensions.

Example

No code snippet is provided as the issue does not contain enough information about the model's implementation or the preprocessing steps.

Notes

The solution may not apply if the issue is not related to image dimensions, but rather to other factors such as image content or model configuration.

Recommendation

Apply a workaround by preprocessing the image to match the expected dimensions, as this is likely to resolve the issue without requiring changes to the model or its configuration.