---
license: apache-2.0
tags:
  - image-editor
  - face-restoration
  - inpainting
  - background-removal
  - upscaling
  - colorization
  - mirror
library_name: pytorch
---

# A.I.M.I Image Editor Tools Mirror

Central mirror for the weights used by the A.I.M.I image editor tab's one-shot AI operations. Re-hosted from upstream for stable URLs; contents unmodified.

## Folders

### `face-restore/` — GFPGAN face restoration
- `GFPGANv1.4.pth` — Tencent GFPGAN v1.4 (Apache 2.0). 332 MB.
- `detection_Resnet50_Final.pth` — facexlib RetinaFace detector (MIT). 104 MB.
- `parsing_parsenet.pth` — facexlib face parser (MIT). 81 MB.

### `removal-lama/` — LAMA object removal
- `big-lama.pt` — LAMA inpainting (Apache 2.0, Samsung Moscow). 196 MB.

### `upscalers/` — ESRGAN + DAT
- `RealESRGAN_x4plus.pth` — Real-ESRGAN x4 (Apache 2.0, Xintao Wang). 64 MB.
- `4xNomos8kDAT.onnx` — DAT x4 photo-fine-tune (CC-BY-4.0, Phips Nomos8k). 61 MB.

### `bgremove-birefnet/` — Background removal
- `birefnet_fp16.onnx` — BiRefNet ONNX (MIT, ZhengPeng7 et al., 2024). 467 MB.

### `colorize-ddcolor/` — B&W colorization
- `ddcolor.onnx` — Alibaba DAMO DDColor (Apache 2.0). 870 MB.

## Licenses

All components permit commercial use. See attributions.

## Attribution

- **GFPGAN**: Tencent ARC (THL A29 Limited / Tencent) — *GFPGAN: Towards Real-World Blind Face Restoration with Generative Facial Prior* (2021). Apache 2.0.
- **facexlib**: Xintao Wang. MIT.
- **LAMA**: Roman Suvorov, Elizaveta Logacheva, Anton Mashikhin, Anastasia Remizova, Arsenii Ashukha, Aleksei Silvestrov, Naejin Kong, Harshith Goka, Kiwoong Park, Victor Lempitsky — *Resolution-robust Large Mask Inpainting with Fourier Convolutions* (WACV 2022). Apache 2.0.
- **Real-ESRGAN**: Xintao Wang, Liangbin Xie, Chao Dong, Ying Shan — *Real-ESRGAN: Training Real-World Blind Super-Resolution with Pure Synthetic Data* (ICCV 2021). Apache 2.0.
- **DAT**: Zheng Chen, Yulun Zhang, Jinjin Gu, Linghe Kong, Xiaokang Yang, Fisher Yu — *Dual Aggregation Transformer Is More Accurate for Super-Resolution* (ICCV 2023). Apache 2.0 code. Photo fine-tune by Phips, released CC-BY-4.0.
- **BiRefNet**: Peng Zheng, Dehong Gao, Deng-Ping Fan, Li Liu, Jorma Laaksonen, Wanli Ouyang, Nicu Sebe — *Bilateral Reference for High-Resolution Dichotomous Image Segmentation* (2024). MIT.
- **DDColor**: Xiaoyang Kang, Tao Yang, Wenqi Ouyang, Peiran Ren, Lingzhi Li, Xuansong Xie (Alibaba DAMO Academy) — *DDColor: Towards Photo-Realistic Image Colorization via Dual Decoders* (ICCV 2023). Apache 2.0.