Text-to-Video
Sana
English
Chinese
image-to-video
SANA
SANA-Video
SANA-Video-2.0
720p
diffusion
LTX-2.3
Instructions to use Efficient-Large-Model/SANA-Video_2.0_5B_720p with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Sana
How to use Efficient-Large-Model/SANA-Video_2.0_5B_720p with Sana:
# Load the model and infer image from text import torch from app.sana_pipeline import SanaPipeline from torchvision.utils import save_image sana = SanaPipeline("configs/sana_config/1024ms/Sana_1600M_img1024.yaml") sana.from_pretrained("hf://Efficient-Large-Model/SANA-Video_2.0_5B_720p") image = sana( prompt='a cyberpunk cat with a neon sign that says "Sana"', height=1024, width=1024, guidance_scale=5.0, pag_guidance_scale=2.0, num_inference_steps=18, ) - Notebooks
- Google Colab
- Kaggle
docs: refresh SANA-Video 2.0 model card
Browse files
README.md
CHANGED
|
@@ -56,15 +56,9 @@ the official inference entry point casts the transformer to BF16.
|
|
| 56 |
- `config.yaml`: matching SANA training and inference configuration
|
| 57 |
- `LICENSE`: Apache License 2.0
|
| 58 |
|
| 59 |
-
Checkpoint SHA256:
|
| 60 |
-
|
| 61 |
-
```text
|
| 62 |
-
7e557554540b4cbbc515166b43a7d307285ab250cedfb48627878227a722d25a
|
| 63 |
-
```
|
| 64 |
-
|
| 65 |
## Verified release example
|
| 66 |
|
| 67 |
-
This sample was generated from the public checkpoint with seed
|
| 68 |
result is 1280 × 736, 193 frames, 24 FPS, and 8.04 seconds long.
|
| 69 |
|
| 70 |
<p align="center">
|
|
@@ -87,14 +81,11 @@ result is 1280 × 736, 193 frames, 24 FPS, and 8.04 seconds long.
|
|
| 87 |
|
| 88 |
## Inference
|
| 89 |
|
| 90 |
-
Support for this checkpoint is
|
| 91 |
-
while [NVlabs/Sana PR #439](https://github.com/NVlabs/Sana/pull/439) is under
|
| 92 |
-
review:
|
| 93 |
|
| 94 |
```bash
|
| 95 |
git clone https://github.com/NVlabs/Sana.git
|
| 96 |
cd Sana
|
| 97 |
-
git checkout release/sana-video-2.0
|
| 98 |
bash environment_setup.sh sana
|
| 99 |
conda activate sana
|
| 100 |
```
|
|
@@ -118,7 +109,7 @@ bash inference_video_scripts/inference_sana_video.sh \
|
|
| 118 |
--step 50 \
|
| 119 |
--fps 24 \
|
| 120 |
--motion_score 20 \
|
| 121 |
-
--seed
|
| 122 |
--work_dir output/sana_video2_t2v_720p_demo
|
| 123 |
```
|
| 124 |
|
|
@@ -176,12 +167,12 @@ appropriate, and add safeguards suited to their application.
|
|
| 176 |
|
| 177 |
- [SANA repository](https://github.com/NVlabs/Sana)
|
| 178 |
- [SANA-Video 2.0 release PR](https://github.com/NVlabs/Sana/pull/439)
|
| 179 |
-
- [SANA-Video 2.0 documentation](https://github.com/NVlabs/Sana/blob/
|
| 180 |
-
- [Model zoo](https://github.com/NVlabs/Sana/blob/
|
| 181 |
- [Verified 5B 720p release video](https://huggingface.co/datasets/Efficient-Large-Model/Sana-assets/resolve/main/Video2/assets/release-demo/sana_video2_5b_720p_rooster.mp4)
|
| 182 |
|
| 183 |
## Citation
|
| 184 |
|
| 185 |
If you use SANA-Video, please cite the SANA-Video work linked from the
|
| 186 |
-
[project page](https://nvlabs.github.io/Sana/
|
| 187 |
citation information will be added when it becomes available.
|
|
|
|
| 56 |
- `config.yaml`: matching SANA training and inference configuration
|
| 57 |
- `LICENSE`: Apache License 2.0
|
| 58 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 59 |
## Verified release example
|
| 60 |
|
| 61 |
+
This sample was generated from the public checkpoint with seed 4. The encoded
|
| 62 |
result is 1280 × 736, 193 frames, 24 FPS, and 8.04 seconds long.
|
| 63 |
|
| 64 |
<p align="center">
|
|
|
|
| 81 |
|
| 82 |
## Inference
|
| 83 |
|
| 84 |
+
Support for this checkpoint is included in the `main` branch of NVlabs/Sana:
|
|
|
|
|
|
|
| 85 |
|
| 86 |
```bash
|
| 87 |
git clone https://github.com/NVlabs/Sana.git
|
| 88 |
cd Sana
|
|
|
|
| 89 |
bash environment_setup.sh sana
|
| 90 |
conda activate sana
|
| 91 |
```
|
|
|
|
| 109 |
--step 50 \
|
| 110 |
--fps 24 \
|
| 111 |
--motion_score 20 \
|
| 112 |
+
--seed 4 \
|
| 113 |
--work_dir output/sana_video2_t2v_720p_demo
|
| 114 |
```
|
| 115 |
|
|
|
|
| 167 |
|
| 168 |
- [SANA repository](https://github.com/NVlabs/Sana)
|
| 169 |
- [SANA-Video 2.0 release PR](https://github.com/NVlabs/Sana/pull/439)
|
| 170 |
+
- [SANA-Video 2.0 documentation](https://github.com/NVlabs/Sana/blob/main/docs/sana_video2.md)
|
| 171 |
+
- [Model zoo](https://github.com/NVlabs/Sana/blob/main/docs/model_zoo.md#sana-video-20)
|
| 172 |
- [Verified 5B 720p release video](https://huggingface.co/datasets/Efficient-Large-Model/Sana-assets/resolve/main/Video2/assets/release-demo/sana_video2_5b_720p_rooster.mp4)
|
| 173 |
|
| 174 |
## Citation
|
| 175 |
|
| 176 |
If you use SANA-Video, please cite the SANA-Video work linked from the
|
| 177 |
+
[project page](https://nvlabs.github.io/Sana/Video2/). SANA-Video 2.0-specific
|
| 178 |
citation information will be added when it becomes available.
|