Thanks for your work. However, when I test your model using LLaVA code, I encounter this problem:
File "/home/zyz//inference_llava.py", line 119, in main
image_tensor = image_processor.preprocess(image, return_tensors='pt')['pixel_values'].half().cuda()
AttributeError: 'NoneType' object has no attribute 'preprocess'
When I load a LLaVA v1.5 model, there is no error. I have checked the "image" is a valid PIL.Image object. Could you please check why there is an AttributeError? Thanks.
Thanks for your work. However, when I test your model using LLaVA code, I encounter this problem:
File "/home/zyz//inference_llava.py", line 119, in main image_tensor = image_processor.preprocess(image, return_tensors='pt')['pixel_values'].half().cuda() AttributeError: 'NoneType' object has no attribute 'preprocess'
When I load a LLaVA v1.5 model, there is no error. I have checked the "image" is a valid PIL.Image object. Could you please check why there is an AttributeError? Thanks.