Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Implementation of Make-A-Video, new SOTA text to video generator from Meta AI, in Pytorch
| Date | Stars |
|---|---|
| 2026-07-24 | 1986 |
| 2026-07-25 | 1986 |
| 2026-07-28 | 1986 |
| 2026-07-30 | 1986 |
| 2026-07-31 | 1987 |
| 2026-08-06 | 1987 |
Today
— stars today
This week
+1 stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.05%/day
<img src="./make-a-video.png" width="400px"></img>
## Make-A-Video - Pytorch (wip)
Implementation of <a href="https://makeavideo.studio/">Make-A-Video</a>, new SOTA text to video generator from Meta AI, in Pytorch. They combine pseudo-3d convolutions (axial convolutions) and temporal attention and show much better temporal fusion.
The pseudo-3d convolutions isn't a new concept. It has been explored before in other contexts, say for protein contact prediction as <a href="https://www.biorxiv.org/content/10.1101/2022.08.04.502748v2.full">"dimensional hybrid residual networks"</a>.
The gist of the paper comes down to, take a SOTA text-to-image model (here they use DALL-E2, but the same learning points would easily apply to Imagen), make a few minor modifications for <a href="https://arxiv.org/abs/2204.03458">attention across time</a> and other ways to skimp on the compute cost, do frame interpolation correctly, get a great video model out.
<a href="https://www.youtube.com/watch?v=AcvmyqGgMh8">AI Coffee Break explanation</a>
## Appreciation
- <a href="https://stability.ai/">Stability.ai</a> for the generous sponsorship to work on cutting edge artificial intelligence research
- <a href="http://www.jonathanho.me/">Jonathan Ho</a> for bringing about a revolution in generative artificial intelligence through <a href="https://arxiv.org/abs/2006.11239">his seminal paper</a>
- <a href="https://github.com/arogozhnikov">Alex</a> for <a href="https://github.com/arogozhnikov/einops">einops</a>, an abstraction that is simply genius. No other word for it.
## Install
```bash
$ pip install make-a-video-pytorch
```
## Usage
Passing in video features
```python
import torch
from make_a_video_pytorch import PseudoConv3d, SpatioTemporalAttention
conv = PseudoConv3d(
dim = 256,
kernel_size = 3
)
attn = SpatioTemporalAttention(
dim = 256,
dim_head = 64,
heads = 8
)
video = torch.randn(1, 256, 8, 16, 16) # (batch, features, frames, height, width)
conv_out = conv(video) # (1, 256, 8, 16, 16)
attn_out = attn(video) # (1, 256, 8, 16, 16)
```
Passing in images (if one were to pretrain on images first), both temporal convolution and attention will be automatically skipped. In other words, you can use this straightforwardly in your 2d Unet and then port it over to a 3d Unet once that phase of the training is done. The temporal modules are initialized to output identity as the paper had done.
```python
import torch
from make_a_video_pytorch import PseudoConv3d, SpatioTemporalAttention
conv = PseudoConv3d(
dim = 256,
kernel_size = 3
)
attn = SpatioTemporalAttention(
dim = 256,
dim_head = 64,
heads = 8
)
images = torch.randn(1, 256, 16, 16) # (batch, features, height, width)
conv_out = conv(images) # (1, 256, 16, 16)
attn_out = attn(images) # (1, 256, 16, 16)
```
You can also control the two modules so that when fed 3-dimensional features, it only does training spatially
```python
import torch
from make_a_video_pytorch import PseudoConv3d, SpatioTemporalAttention
conv = PseudoConv3d(
dim = 256,
kernel_size = 3
)
attn = SpatioTemporalAttention(
dim = 256,
dim_head = 64,
heads = 8
)
video = torch.randn(1, 256, 8, 16, 16) # (batch, features, frames, height, width)
# below it will not train across time
conv_out = conv(video, enable_time = False) # (1, 256, 8, 16, 16)
attn_out = attn(video, enable_time = False) # (1, 256, 8, 16, 16)
```
Full `SpaceTimeUnet` that is agnostic to images or video training, and where even if video is passed in, time can be ignored
```python
import torch
from make_a_video_pytorch import SpaceTimeUnet
unet = SpaceTimeUnet(
dim = 64,
channels = 3,
dim_mult = (1, 2, 4, 8),
resnet_block_depths = (1, 1, 1, 2),
temporal_compression = (False, False, False, True),
self_attns = (False, False, False, True),
condition_on_timestep = False,
attn_pos_bias = False,
flash_attn = True
).cuda()
# train on images
images = torch.Excerpt of 6,300 characters
Read on GitHub
Phil Wang · United States
41
1
Would you bet a product on this? Bounded 0–100 and slow moving.
matched fp:507c8a50d88d88a1, topic:text-to-video, desc:text-to-video, readme:text-to-video
matched fp:507c8a50d88d88a1, topic:deep-learning