| --- |
| title: Between Frames Video Generator |
| emoji: 🎞️ |
| colorFrom: yellow |
| colorTo: red |
| sdk: gradio |
| sdk_version: 6.20.0 |
| app_file: app.py |
| pinned: false |
| license: apache-2.0 |
| startup_duration_timeout: 1h |
| models: |
| - Wan-AI/Wan2.1-FLF2V-14B-720P-diffusers |
| preload_from_hub: |
| - Wan-AI/Wan2.1-FLF2V-14B-720P-diffusers |
| --- |
| |
| # Between Frames |
|
|
| A Hugging Face Space for generating a video from: |
|
|
| - a required start image; |
| - an optional end image; and |
| - a text prompt describing the action, subject motion, and camera movement; |
| - a selectable duration of 1, 2, or 3 seconds. |
|
|
| The app uses [`Wan-AI/Wan2.1-FLF2V-14B-720P-diffusers`](https://huggingface.co/Wan-AI/Wan2.1-FLF2V-14B-720P-diffusers). When no end image is supplied, the same pipeline runs in regular image-to-video mode. |
|
|
| ## Deploy |
|
|
| 1. Create a new **Gradio** Space on Hugging Face. |
| 2. Upload the files in this repository. |
| 3. Select **ZeroGPU** in **Settings → Hardware**. |
| 4. Wait for the model weights to download on the first build/request. |
|
|
| The model is large. CPU-only hardware is not supported. The app requests ZeroGPU's 96 GB `xlarge` tier while a generation is running and dynamically scales the reservation with the selected duration, resolution, and step count. Usage remains subject to Hugging Face's daily ZeroGPU quota and queue. |
|
|
