Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,127 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# FaceFusion 3.0.0 – Complete ONNX Model Repository
|
| 2 |
+
|
| 3 |
+
This repository contains the **full, unmodified ONNX model set** duplicated from
|
| 4 |
+
`facefusion/models-3.0.0`.
|
| 5 |
+
|
| 6 |
+
All models are provided **for inference only** and are intended for use with **FaceFusion**
|
| 7 |
+
or custom ONNX Runtime pipelines.
|
| 8 |
+
|
| 9 |
+
Each `.onnx` file is paired with a `.hash` file used for **integrity verification**.
|
| 10 |
+
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
## General Information
|
| 14 |
+
|
| 15 |
+
- **Framework:** ONNX
|
| 16 |
+
- **Origin:** facefusion/models-3.0.0
|
| 17 |
+
- **Status:** Inference-ready
|
| 18 |
+
- **Execution:** CPU / CUDA / TensorRT / DirectML / CoreML (provider dependent)
|
| 19 |
+
- **Training:** Not supported
|
| 20 |
+
|
| 21 |
+
---
|
| 22 |
+
|
| 23 |
+
## Complete Model List
|
| 24 |
+
|
| 25 |
+
### Landmark Detection & Alignment
|
| 26 |
+
- `2dfan4.onnx`
|
| 27 |
+
- `fan_68_5.onnx`
|
| 28 |
+
|
| 29 |
+
### ArcFace / Identity Conversion
|
| 30 |
+
- `arcface_w600k_r50.onnx`
|
| 31 |
+
- `arcface_converter_ghost.onnx`
|
| 32 |
+
- `arcface_converter_simswap.onnx`
|
| 33 |
+
|
| 34 |
+
### Face Detection
|
| 35 |
+
- `retinaface_10g.onnx`
|
| 36 |
+
- `scrfd_2.5g.onnx`
|
| 37 |
+
- `yoloface_8n.onnx`
|
| 38 |
+
|
| 39 |
+
### Face Segmentation
|
| 40 |
+
- `bisenet_resnet_34.onnx`
|
| 41 |
+
- `dfl_xseg.onnx`
|
| 42 |
+
|
| 43 |
+
### Face Swapping
|
| 44 |
+
- `inswapper_128.onnx`
|
| 45 |
+
- `inswapper_128_fp16.onnx`
|
| 46 |
+
- `simswap_256.onnx`
|
| 47 |
+
- `simswap_unofficial_512.onnx`
|
| 48 |
+
- `uniface_256.onnx`
|
| 49 |
+
- `blendswap_256.onnx`
|
| 50 |
+
- `ghost_1_256.onnx`
|
| 51 |
+
- `ghost_2_256.onnx`
|
| 52 |
+
- `ghost_3_256.onnx`
|
| 53 |
+
|
| 54 |
+
### Face Restoration & Enhancement
|
| 55 |
+
- `gfpgan_1.2.onnx`
|
| 56 |
+
- `gfpgan_1.3.onnx`
|
| 57 |
+
- `gfpgan_1.4.onnx`
|
| 58 |
+
- `gpen_bfr_256.onnx`
|
| 59 |
+
- `gpen_bfr_512.onnx`
|
| 60 |
+
- `gpen_bfr_1024.onnx`
|
| 61 |
+
- `gpen_bfr_2048.onnx`
|
| 62 |
+
- `restoreformer_plus_plus.onnx`
|
| 63 |
+
- `codeformer.onnx`
|
| 64 |
+
|
| 65 |
+
### Face Attributes / Analysis
|
| 66 |
+
- `fairface.onnx`
|
| 67 |
+
- `styleganex_age.onnx`
|
| 68 |
+
|
| 69 |
+
### Live Portrait / Motion Transfer
|
| 70 |
+
- `live_portrait_feature_extractor.onnx`
|
| 71 |
+
- `live_portrait_motion_extractor.onnx`
|
| 72 |
+
- `live_portrait_eye_retargeter.onnx`
|
| 73 |
+
- `live_portrait_lip_retargeter.onnx`
|
| 74 |
+
- `live_portrait_generator.onnx`
|
| 75 |
+
- `live_portrait_stitcher.onnx`
|
| 76 |
+
|
| 77 |
+
### Lip Synchronization
|
| 78 |
+
- `wav2lip_96.onnx`
|
| 79 |
+
- `wav2lip_gan_96.onnx`
|
| 80 |
+
|
| 81 |
+
### Super-Resolution & Upscaling
|
| 82 |
+
- `real_esrgan_x2.onnx`
|
| 83 |
+
- `real_esrgan_x2_fp16.onnx`
|
| 84 |
+
- `real_esrgan_x4.onnx`
|
| 85 |
+
- `real_esrgan_x4_fp16.onnx`
|
| 86 |
+
- `real_esrgan_x8.onnx`
|
| 87 |
+
- `real_esrgan_x8_fp16.onnx`
|
| 88 |
+
- `real_hatgan_x4.onnx`
|
| 89 |
+
- `ultra_sharp_x4.onnx`
|
| 90 |
+
- `lsdir_x4.onnx`
|
| 91 |
+
- `nomos8k_sc_x4.onnx`
|
| 92 |
+
- `span_kendata_x4.onnx`
|
| 93 |
+
- `clear_reality_x4.onnx`
|
| 94 |
+
|
| 95 |
+
### Colorization
|
| 96 |
+
- `deoldify.onnx`
|
| 97 |
+
- `deoldify_artistic.onnx`
|
| 98 |
+
- `deoldify_stable.onnx`
|
| 99 |
+
- `ddcolor.onnx`
|
| 100 |
+
- `ddcolor_artistic.onnx`
|
| 101 |
+
|
| 102 |
+
### Audio / Classification / Misc
|
| 103 |
+
- `kim_vocal_2.onnx`
|
| 104 |
+
- `open_nsfw.onnx`
|
| 105 |
+
- `peppa_wutz.onnx`
|
| 106 |
+
|
| 107 |
+
---
|
| 108 |
+
|
| 109 |
+
## Hash Files
|
| 110 |
+
|
| 111 |
+
Every model includes a corresponding `.hash` file:
|
| 112 |
+
|
| 113 |
+
- Ensures file integrity
|
| 114 |
+
- Required by FaceFusion
|
| 115 |
+
- Must not be removed or modified
|
| 116 |
+
|
| 117 |
+
---
|
| 118 |
+
|
| 119 |
+
## Example Usage
|
| 120 |
+
|
| 121 |
+
```python
|
| 122 |
+
import onnxruntime as ort
|
| 123 |
+
|
| 124 |
+
session = ort.InferenceSession(
|
| 125 |
+
"inswapper_128.onnx",
|
| 126 |
+
providers=["CUDAExecutionProvider", "CPUExecutionProvider"]
|
| 127 |
+
)
|