Yanyu

Thanks for pointing this out. We will revise the manuscript and add ablation results on LN. Initially, GN with `groups=1`, which is the case in current ablation, is supposed to...

Thanks for your interest in our work. Good news, now our model definition and pretrained models are in collections of [timm](https://github.com/rwightman/pytorch-image-models).

Thanks for your interest in our work. We are going to add the following speed results in the final manuscript: 1. Speed on NVIDIA A100, deployed with TensorRT 2. Speed...

I'm not quite certain about this issue. Perhaps 1. It is because the CPU version of PyTorch, 2. It is because some packages are not properly installed (such as timm)...

It seems that you are using a new script `Demo.py`, did you import the EfficientFormer models in that script? Similar to `main.py`: ``` from models import * ```

I can reproduce the bug with latest PyTorch. ``` WARNING:root:Torch version 1.13.0a0+08820cb has not been tested with coremltools. You may run into unexpected errors. Torch 1.10.2 is the most recent...

Hi Shudeng, yes, the root cause of the rigid resolution problem comes from position embeddings in ViT, which occur as attention bias in our model. If you want to try...

Hi, could you provide more details? 1. Is the pretrained model successfully (and fully) loaded? Maybe you can first verify mIoU on ADE20K to check. 2. If so, have you...

Hi, Thanks for your reply. I think it can be reproduced through the official example here, https://huggingface.co/docs/diffusers/en/using-diffusers/svd#stable-video-diffusion by setting `pipeline(..., num_videos_per_prompt=2, ...)`

Hi, In most of our experiments, we use 16gpu x 128bs to 16gpu x 256bs for training. We do notice that it is not easy to achieve best results when...