Spaces:
Build error
which model exactly is used in this demo?
the website gives me really impressive results. while I'm trying to run the prediction locally with github code and models, the result was different. https://github.com/salesforce/BLIP/blob/main/demo.ipynb
i'm wondering what is the model used here? I tried model_base_capfilt_large.pth. thanks
You can see the model used by the demo here (https://huggingface.co/spaces/Salesforce/BLIP/blob/main/app.py#L24). This is the BLIP_large model, finetuned on COCO.
Although you may find the results different, I'd expect captions from model_base_capfilt_large also make sense. Thanks.
Hi, I have try BLIP_large model, which finetuned on COCO, but it seems only generate about 10 words caption. Is there any sulotion to generate more detail caption.
@mengmingming Try BLIP-2 and InstructBLIP (https://github.com/salesforce/LAVIS/tree/main/projects)