My Implementation of " Structure and Content-Guided Video Synthesis with Diffusion Models" by RunwayML. "Input videos x are encoded to z0 with a fixed encoder E and diffused to zt. We extract a structure representation s by encoding depth maps obtained with MiDaS, and a content representation c by encoding one of the frames with CLIP. The model then learns to reverse the diffusion process in the latent space, with the help of s, which gets concatenated to zt, as well as c, which is provided via cross-attention blocks. During inference (right), the structure s of an input video is provided in the same manner. To specify content via text, we convert CLIP text embeddings to image embeddings via a prior."
pip3 install gen1
import torch
from gen1.model import Gen1
# Create an instance of the Gen1 model
model = Gen1()
# Generate random input images and video tensors
images = torch.randn(1, 3, 128, 128)
video = torch.randn(1, 3, 16, 128, 128)
# Pass the input images and video through the model's forward method
run_out = model.forward(images, video)
Here is a summary table of the datasets used in the Structure and Content-Guided Video Synthesis with Diffusion Models paper:
Dataset | Type | Size | Domain | Description | Source |
---|---|---|---|---|---|
Internal dataset | Images | 240M | General | Uncaptioned images | Private |
Custom video dataset | Videos | 6.4M clips | General | Uncaptioned short video clips | Private |
DAVIS | Videos | - | General | Video object segmentation | Link |
Stock footage | Videos | - | General | Diverse video clips | - |
@misc{2302.03011,
Author = {Patrick Esser and Johnathan Chiu and Parmida Atighehchian and Jonathan Granskog and Anastasis Germanidis},
Title = {Structure and Content-Guided Video Synthesis with Diffusion Models},
Year = {2023},
Eprint = {arXiv:2302.03011},