Spaces:
Running
Running
Update GhostPackDemo/readme.md
Browse files- GhostPackDemo/readme.md +17 -16
GhostPackDemo/readme.md
CHANGED
|
@@ -1,24 +1,25 @@
|
|
| 1 |
-
|
|
|
|
| 2 |
|
| 3 |
-
|
| 4 |
|
| 5 |
-
>
|
| 6 |
-
> Powered by **HunyuanVideo**, wrapped in an easy Gradio UI.
|
| 7 |
|
| 8 |
---
|
| 9 |
|
| 10 |
-
##
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
|
| 14 |
-
|
| 15 |
-
|
| 16 |
-
| VRAM-Aware Loading | Automatic model slicing/offloading for low-VRAM (< 60 GB) GPUs. |
|
| 17 |
-
| Offline Model Cache | All models downloaded once to `hf_download/` for fast, private reuse. |
|
| 18 |
-
| One-Command Installer | `./install.py` sets up **Python venv**, **CUDA** check, **HF login**, **model fetch**. |
|
| 19 |
-
| Cross-Platform | Works on Linux π§ & Windows πͺ (WSL / PowerShell). |
|
| 20 |
|
| 21 |
---
|
| 22 |
|
| 23 |
-
##
|
| 24 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
<!-- README.md -->
|
| 2 |
+
# π»π¬ **GhostPackDemo** β Local AI Image-to-Video Generator
|
| 3 |
|
| 4 |
+
 
|
| 5 |
|
| 6 |
+
> Turn a single image into a cinematic video clip with one command.
|
|
|
|
| 7 |
|
| 8 |
---
|
| 9 |
|
| 10 |
+
## β¨ Features
|
| 11 |
+
- **Image β Video diffusion** using HunyuanVideo & FramePack.
|
| 12 |
+
- **Progress preview**: live denoise frames + progress bar.
|
| 13 |
+
- **Low-VRAM mode** (< 60 GB) with automatic model swapping.
|
| 14 |
+
- **Offline cache** in `hf_download/`; no repeated downloads.
|
| 15 |
+
- **One-click installer** `./install.py`.
|
|
|
|
|
|
|
|
|
|
|
|
|
| 16 |
|
| 17 |
---
|
| 18 |
|
| 19 |
+
## π Quickstart
|
| 20 |
+
```bash
|
| 21 |
+
git clone https://huggingface.co/spaces/ghostai1/GhostPack.git
|
| 22 |
+
cd GhostPack/GhostPackDemo
|
| 23 |
+
./install.py # create venv, install deps, cache model
|
| 24 |
+
source venv/bin/activate # Windows: venv\Scripts\Activate.ps1
|
| 25 |
+
python ghostpack_main_fixed.py --server 0.0.0.0 --port 7860 --inbrowser
|