---
license: mit
language:
- en
tags:
- animation
- stroke
- interpolation
- 2D
- image
- video
---
# Stroke Interpolation Model

To read the paper: https://drive.google.com/file/d/1EESd81NSs93OJYb42DartC5udTlOShRp/view?usp=sharing 

## Example

![image/gif](https://cdn-uploads.huggingface.co/production/uploads/62d36d11274bf1ef84f61d66/nY0uCIdQkWOOAzvL_Gj9e.gif)
The model predicts the inbetween frames (mid frame), given two key frames.
## Installation
```
pip install opencv-python 
pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu128
```

## Run the Program
```
py combined.py
```
This will load the model from checkpoints.

## Evaluate
```
py generate_eval.py
```
This generates images for evaluation.

```
py eval.py
```
This evaluates the generated images.

## Test
```
py test.py
```
Add frames in test_frames folder.

```
py video.py
```
This combines those 3 frames into .mp4 format.

## Dataset
Dataset is available at: [Google Drive Link](https://drive.google.com/file/d/1vyu_ePFN9sFjqxc-sPdSWuSCLnWFVUT7/view?usp=sharing)