p1atdev / stable-diffusion-webui-blip2-captioner Goto Github PK

Where does the Coco model get saved to? I can't find it in either the master/model folder or the .cache\huggingface\hub one.

Curious if this is working in AUTOMATIC1111 v1.5? I installed the extension but I'm not seeing a new tab anywhere.

Hello / Good evening
I encounter an anomaly after the installation and reboot of stable diffusion, this one tells me to install a new version of pip, but it is already present and installed, multiple errors that then come in sequence,

I tried Single, it's fine.
But when I tried Batch, it showed the following error.

Error: Sizes of tensors must match except in dimension 1. Expected size 9 but got size 3 for tensor number 1 in the list.

What should I do?

error on batch run but OK in single image
error message says:

Error: Sizes of tensors must match except in dimension 1. Expected size 49 but got size 7 for tensor number 1 in the list.

Any idea?

Hi, I'm wondering if there is a better model than coco to get a higher quality of clip?

When I first clicked to Interrogate I see 2 loading and after it loaded it generate text 2 times :(
Loading coco model...
Loading coco model...
Position interpolate from 16x16 to 26x26
Position interpolate from 16x16 to 26x26
Model coco loaded
Generating captions...
a man in front of many women dressed in military uniforms
Model coco loaded
Generating captions...
kim yong bin poses for a photograph with women wearing hats

If unload model and click again - all fine!
It took X2 more time on first click :((

Add the ability to generate more captions in one click.

num_captions=,

Make it so that when batch processing, save to one csv file
img.jpg|caption

Thank you very much for your work!

Hello author!
May I ask if your project can generate text descriptions for images in the image dataset in batches?