BAAI
/

File size: 4,070 Bytes
b6f77bf
 
 
 
 
 
 
 
 
 
 
 
1e66007
b6f77bf
 
 
1e66007
b6f77bf
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
9487909
 
 
 
 
 
 
 
 
b6f77bf
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
---
license: apache-2.0
tags:
  - text-to-video
  - video-generation
  - baai-nova
---

# NOVA (d48w1024-osp480) Model Card 

## Model Details
- **Developed by:** BAAI
- **Model type:** Non-quantized Autoregressive Text-to-Video Generation Model
- **Model size:** 645M
- **Model precision:** torch.float16 (FP16)
- **Model resolution:** 768x480
- **Model Description:** This is a model that can be used to generate and modify videos based on text prompts. It is a [Non-quantized Video Autoregressive (NOVA)](https://arxiv.org/abs/2412.14169) diffusion model that uses a pretrained text encoder ([Phi-2](https://huggingface.co/microsoft/phi-2)) and one VAE video tokenizer ([OpenSoraPlanV1.2-VAE](https://huggingface.co/LanguageBind/Open-Sora-Plan-v1.2.0)).
- **Model License:** [Apache 2.0 License](LICENSE)
- **Resources for more information:** [GitHub Repository](https://github.com/baaivision/NOVA).

## Examples

Using the [🤗's Diffusers library](https://github.com/huggingface/diffusers) to run NOVA in a simple and efficient manner.

```bash
pip install diffusers transformers accelerate imageio[ffmpeg]
pip install git+ssh://git@github.com/baaivision/NOVA.git
```

Running the pipeline:

```python
import torch
from diffnext.pipelines import NOVAPipeline
from diffnext.utils import export_to_image, export_to_video

model_id = "BAAI/nova-d48w1024-osp480"
model_args = {"torch_dtype": torch.float16, "trust_remote_code": True}
pipe = NOVAPipeline.from_pretrained(model_id, **model_args)
pipe = pipe.to("cuda")

prompt = "Many spotted jellyfish pulsating under water."

image = pipe(prompt, max_latent_length=1).frames[0, 0]
export_to_image(image, "jellyfish.jpg")

video = pipe(prompt, max_latent_length=9).frames[0]
export_to_video(video, "jellyfish.mp4", fps=12)

# Increase AR and diffusion steps for better video quality.
video = pipe(
  prompt,
  max_latent_length=9,
  num_inference_steps=128,  # default: 64
  num_diffusion_steps=100,  # default: 25
).frames[0]
export_to_video(video, "jellyfish_v2.mp4", fps=12)
```

# Uses

## Direct Use 
The model is intended for research purposes only. Possible research areas and tasks include

- Research on generative models.
- Applications in educational or creative tools.
- Generation of artworks and use in design and other artistic processes.
- Probing and understanding the limitations and biases of generative models.
- Safe deployment of models which have the potential to generate harmful content.

Excluded uses are described below.

#### Out-of-Scope Use
The model was not trained to be factual or true representations of people or events, and therefore using the model to generate such content is out-of-scope for the abilities of this model.

#### Misuse and Malicious Use
Using the model to generate content that is cruel to individuals is a misuse of this model. This includes, but is not limited to:

- Mis- and disinformation.
- Representations of egregious violence and gore.
- Impersonating individuals without their consent.
- Sexual content without consent of the people who might see it.
- Sharing of copyrighted or licensed material in violation of its terms of use.
- Intentionally promoting or propagating discriminatory content or harmful stereotypes.
- Sharing content that is an alteration of copyrighted or licensed material in violation of its terms of use.
- Generating demeaning, dehumanizing, or otherwise harmful representations of people or their environments, cultures, religions, etc.

## Limitations and Bias

### Limitations

- The autoencoding part of the model is lossy.
- The model cannot render complex legible text.
- The model does not achieve perfect photorealism.
- The fingers, .etc in general may not be generated properly.
- The model was trained on a subset of the web datasets [LAION-5B](https://laion.ai/blog/laion-5b/) and [COYO-700M](https://github.com/kakaobrain/coyo-dataset), which contains adult, violent and sexual content.

### Bias
While the capabilities of image generation models are impressive, they can also reinforce or exacerbate social biases.