cpystan
cpystan
Thank you for your interest in our work. In terms of the weights, we will open one of our weights this week. In terms of the processing procedures, please give...
have you included the following code? resume_path = os.path.join(args.checkpoint_dir, 'model_best.pth') print("Loading checkpoint: {} ...".format(resume_path)) checkpoint = torch.load(resume_path)['state_dict'] model_dict = model.state_dict() state_dict = {k:v for k,v in checkpoint.items()} model_dict.update(state_dict) model.load_state_dict(model_dict)
> Hello, thank you for providing the repo. However, I am having problems for using the ResNet based [ckpt](https://drive.google.com/file/d/1BDT345Jh9iQWjaLyeWm_ioiOibxFwcR3/view?usp=sharing) you have provided. The state_dict is different with the r2gen model...
https://github.com/cpystan/WSI-VQA/blob/master/downstream/show.ipynb you might refer to this file.
PathText.json文件里应该是都包含了的。ocr/datset_csv里的文件列出了对应的文件名。
Thank you for your suggestion. We are considering to release the text generation code if it would be helpful for you.
I have the same problem. Have you solved this problem now? If you could give me some suggestions, I would be very grateful!
Good Question. We use the slides whose names own 'DX' (which means the diagnostic slide). And for the situation that one patient has several DX slides, we use the 'DX1'...
I will check the code. Have you tried with our provided checkpoint? Does it still fail?
That means your ckpt and my ckpt neither can generate correct answer?