Stable Video Diffusion Training Code 🚀

We have attempted to incorporate layout control on top of img2video, which makes the motion of objects more controllable, similar to what is demonstrated in the image below. The code and weights will be updated soon.

Init Image	Gen Video

size=(512, 320), motion_bucket_id=127, fps=7, noise_aug_strength=0.00
generator=torch.manual_seed(111)

Init Image	Before Fine-tuning	After Fine-tuning

Note that BDD100K is a driving video/image dataset, but this is not a necessity for training. Any video can be used to initiate your training. Please refer to the DummyDataset data reading logic. In short, you only need to modify self.base_folder. Then arrange your videos in the following file structure:

self.base_folder
    ├── video_name1
    │   ├── video_frame1
    │   ├── video_frame2
    │   ...
    ├── video_name2
    │   ├── video_frame1
        ├── ...

This training configuration is for reference only, I set all parameters of unet to be trainable during the training and adopted a learning rate of 1e-5.

accelerate launch train_svd.py \
    --pretrained_model_name_or_path=/path/to/weight \
    --per_gpu_batch_size=1 --gradient_accumulation_steps=1 \
    --max_train_steps=50000 \
    --width=512 \
    --height=320 \
    --checkpointing_steps=1000 --checkpoints_total_limit=1 \
    --learning_rate=1e-5 --lr_warmup_steps=0 \
    --seed=123 \
    --mixed_precision="fp16" \
    --validation_steps=200

While the codebase is functional and provides an enhancement in video generation(maybe? 🤷), it's important to note that there are still some uncertainties regarding the finer details of its implementation.

• Support text2video (WIP)
• Support more conditional inputs, such as layout

Feel free to fork this repository, submit pull requests, or open issues to discuss potential changes or report bugs. With your valuable input, we can continuously improve SVD_Xtend for the community.