Yuxuan Zhang

We recommend using the fine-tuning code provided by the diffusers version, which we will release in early October. This issue will be closed as it cannot be reproduced

We are still testing the relevant code together with the Huggingface team, such as https://github.com/huggingface/diffusers/pull/9482 Currently, the fine-tuned videos include both normal and faulty ones, and we are verifying the...

In the fine-tuning section, we will provide a fine-tuning version of diffusers. The model version of diffusers will be released this week, and it is expected to be fine-tuned on...

ofs is a constant, mainly adding a constant to the embed of the model

What GPU are you using, it shouldn't be this slow. Also, the video should be 6 seconds long, can you calculate how long the average step took?

This speed is clearly incorrect, however, for equipment like yours, I suggest operating according to this plan This will significantly increase the speed

This code is correct, I did not see any errors ``` video_pt = pipe_image( image=image, prompt=prompt, negative_prompt=negative_prompt, num_inference_steps=50, num_videos_per_prompt=1, use_dynamic_cfg=True, output_type="pt", guidance scale 7.0 number of frames 49 generator=torch.Generator(device="cuda").manual_seed(seed), ).frames[0]...

This is clearly not the level of the A6000, even the T4 is faster than this

I use A100 for 180 seconds with the 5B model

For 4090, you can completely remove ``` pipe_image.enable_sequential_cpu_offload() ``` and just move pipe.to("cuda"), should work Currently, there is indeed no way to visualize the intermediate results