hht2001

hello, I wonder can Pytorch_Android support Kotlin? :)

the pretrained-layer's param has been re-init

5

hi, I use the 'from_pretrain ' func to load the pretrain model ,but I found the linear param will be re-init when I simply replace the nn.Linear with lora.Linear

what's the difference between the 'evaluate' method and 'evaluate_raw' method in the class 'VQADataset' in 'vqa_data.py'?

the original_size and target_size in the train process

3

Hi, great work I want to repeat it in diffusers in diffusers, what should the parameter original_size and target_size embedding be? thanks

`retry = 3 while retry > 0: try: img_bytes = self.file_client.get(gt_path, 'gt') except (IOError, OSError) as e: logger = get_root_logger() logger.warn(f'File client error: {e}, remaining retry times: {retry - 1}')...

No such file or directory: '/root/VL_adapter/datasets/vqa/train.json'

1

how to solve it...there is no 'train.json' file in the download link

 在计算M时采用L2范数的平方，理应为各元素的平方和 但是在codes/models/archs/TCVC_IDC_arch.py. line 215中为和的平方 不知道对结果有什么影响？

Bug of PixArtAlphaPipeline

1

### Describe the bug When I use with torch.autocast("cuda"): images = pipe(...) the output is noise  ### Reproduction import torch from diffusers import PixArtAlphaPipeline pipe = PixArtAlphaPipeline.from_pretrained("PixArt-alpha/PixArt-XL-2-1024-MS", torch_dtype=torch.float16) pipe.enable_model_cpu_offload()...

Hi, I found that u said there is a "sudden converge" I think that because of the uniform sampling of timestep in training. If use (large timestep, 1000) in the...