File size: 2,098 Bytes
e5e3587
 
 
 
 
abdd865
 
 
e5e3587
 
 
 
 
 
 
 
 
eb214b7
 
e5e3587
eb214b7
e5e3587
 
eb214b7
 
e5e3587
 
eb214b7
e5e3587
 
eb214b7
e5e3587
 
 
 
 
 
 
 
 
 
 
 
 
eb214b7
 
 
 
 
 
e5e3587
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
d9904e6
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
---
license: apache-2.0
library_name: diffusers
pipeline_tag: image-to-image
tags:
- optical-flow prediction
- motion prediction
- diffusion
---

# FOFPred: Language-Driven Future Optical Flow Prediction

**FOFPred** is a diffusion-based model that predicts future optical flow from a single image guided by natural language instructions. Given an input image and a text prompt describing a desired action (e.g., *"Moving the water bottle from right to left"*), FOFPred generates 4 sequential optical flow frames showing how objects would move.

## Usage

```python
import einops
import numpy as np
import torch
from diffusers import DiffusionPipeline
from PIL import Image

# Load pipeline with trust_remote_code
pipeline = DiffusionPipeline.from_pretrained(
    "Salesforce/FOFPred",
    torch_dtype=torch.bfloat16,
    trust_remote_code=True,
).to("cuda")

# Run inference
results = pipeline(
    prompt="Moving the water bottle from right to left.",
    input_images=[Image.open("your_image.jpg")],
    width=256,
    height=256,
    num_inference_steps=1,
    num_images_per_prompt=4,
    frame_count=4,
    generator=torch.Generator(device="cuda").manual_seed(42),
    output_type="pt",
)

flow_frames = results.images  # [B, F, C, H, W]

output_tensor = flow_frames[0]  # [F, C, H, W]
output_np = pipeline.image_processor.pt_to_numpy(output_tensor)  # [F, H, W, C]
reshaped = einops.rearrange(output_np, "f h w c -> h (f w) c")
img = Image.fromarray((reshaped * 255).astype(np.uint8))
img.save("output_combined.png")
```

## Architecture

| Component | Model |
|-----------|-------|
| **V-LLM** | Qwen2.5-VL-3B-Instruct |
| **DiT** | OmniGen2Transformer3DModel |
| **VAE** | FLUX.1-dev AutoencoderKL |
| **Scheduler** | FlowMatchEulerDiscreteScheduler |

## Acknowledgements

- [OmniGen2](https://github.com/VectorSpaceLab/OmniGen2)
- [Qwen2.5-VL](https://huggingface.co/Qwen/Qwen2.5-VL)
- [Flux VAE](https://huggingface.co/black-forest-labs/FLUX.1-dev)

## License

Our code and weights are released under the [CC by-NC 4.0 license](https://creativecommons.org/licenses/by-nc/4.0/deed.en).