File size: 1,873 Bytes
d8a0391 bd3accc d8a0391 ad2dfd2 d8a0391 8329186 d8a0391 |
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 |
---
license: mit
pipeline_tag: video-to-video
library_name: pytorch
tags:
- computer-vision
- video
- video-frame-interpolation
- vfi
- video-to-video
- comfyui
- pytorch
---
# SnJake Sapsan-VFI
Sapsan-VFI is a **x2 frame interpolation** model for video. It inserts a single middle frame between every input pair, effectively doubling the FPS.
## Examples
<video controls autoplay loop muted playsinline src="https://huggingface.co/SnJake/Sapsan-VFI/resolve/main/videos/1_comparison_x2.mp4" width="100%"></video>
<video controls autoplay loop muted playsinline src="https://huggingface.co/SnJake/Sapsan-VFI/resolve/main/videos/2_comparison_x2.mp4" width="100%"></video>
<video controls autoplay loop muted playsinline src="https://huggingface.co/SnJake/Sapsan-VFI/resolve/main/videos/3_comparison_x2.mp4" width="100%"></video>
## How to use in ComfyUI
The model is designed to work with the **Sapsan-VFI** ComfyUI node.
1. Install the node from the [GitHub repo](https://github.com/SnJake/SnJake_Sapsan-VFI).
2. Download the weights from this repository.
3. Place the file(s) into `ComfyUI/models/sapsan_vfi/`.
4. Select the weights in the node dropdown and run the workflow.
Recommended workflow:
Example workflow can be found in `Example Workflow` folder in [GitHub repo](https://github.com/SnJake/SnJake_Sapsan-VFI).
Notes:
- The node has a `console_progress` toggle to print progress in the ComfyUI console.
## Weights
- `Sapsan-VFI.safetensors`
- `Sapsan-VFI.pt`
## Training Details
- Created out of curiosity and personal interest.
- Total epochs: **11**
- Dataset: **2700 videos**
- Shards: **151** shards of **1000** shadrs in each. 151 000 triplets.
Training code is included in `training_code/` for reference.
## Disclaimer
This project was made purely for curiosity and personal interest. The code was written by GPT-5.2 Codex.
|