"LLaVA is not available. Please add text manually."

[Sumsar3]

I'm trying to get LLaVA to work in SUPIR but I'm not able to do it. I downloaded the folders (143GB lol) and move it to the models folder:

Then I set the paths on CKPT_PTH.py:

I run gradio with python gradio_demo.py --use_tile_vae --use_image_slider --loading_half_params --load_8bit_llava

When I'm into gradio and I click on "LLaVA run" I get this error on Prompt Box:

No errors on cmd window.

What I'm doing wrong? Thanks.

[Fanghua-Yu]

This prompt is occurred when use_llava = False:

SUPIR/gradio_demo.py

Lines 75 to 83 in bca9a72

 def llave_process(input_image, temperature, top_p, qs=None): 
     torch.cuda.set_device(LLaVA_device) 
     if use_llava: 
         LQ = HWC3(input_image) 
         LQ = Image.fromarray(LQ.astype('uint8')) 
         captions = llava_agent.gen_image_caption([LQ], temperature=temperature, top_p=top_p, qs=qs) 
     else: 
         captions = ['LLaVA is not available. Please add text manually.'] 
     return captions[0] 

It should be true with your gradio setting. Try to find something wrong with this variant. (sync the repo first)

[Sumsar3]

Closing the issue since was a gradio problem. Thanks!