FeiiYin

Hi, I made the code run via change the line 61 in `deepfloyd_prompt_processor.py` to ``` text_encoder = T5EncoderModel.from_pretrained( pretrained_model_name_or_path, subfolder="text_encoder", torch_dtype=torch.float16, # suppress warning load_in_8bit=True, variant="8bit", device_map="auto", ) ``` I...

Thanks for your attention. That is kind of an artifact of this framework. It mainly comes from the introduction of a post-training calibration network. The artifact will be alleviated a...