google-research / magvit

Official JAX implementation of MAGVIT: Masked Generative Video Transformer

https://magvit.cs.cmu.edu

google-research/magvit Issues

jax Resnet50 pretrain model for perceptual loss
Updated 24 days ago1
Question about the training loss
Updated 24 days ago
Training 2D VQGAN on UCF 101
Updated a month ago1
Models
Updated a month ago2
discriminator requires temporal frames of 32 but data contains 16 frames
Updated 2 months ago
SPAE code release
Updated 2 months ago7
Update: Model Weight Release Not Approved
Updated 3 months ago1
How do export magvit tokens and embeddings?
Updated 4 months ago
Which model is better for perceptual loss， vgg19 or resnet50?
Updated 4 months ago
Great work, looking forward to model release, testing
Updated 4 months ago1
7 months passed, the weights is still unavailable...
Updated 4 months ago1
SPAE code release
Updated 5 months ago6
Is there a de-tokenize process?
Updated 7 months ago
weights
Updated 8 months ago
Model weights and loading instructions are coming soon.
Updated 8 months ago7
Google Colab
Updated 8 months ago2
About ImageNet Pretraining
Updated 10 months ago
New
Updated a year ago
SPAE Project codes
Updated a year ago1
Code release timeline
Closed a year ago11
Can anybody believe this project got SOTA in FVD?
Closed a year ago2
you should make a hugging face account.
Updated a year ago