The model is available via the official GPEN GitHub repository and its associated ModelScope.
The "gpen-bfr-2048.pth" model could be used for various applications, including: gpen-bfr-2048.pth
It can be combined with other background restoration tools (like Real-ESRGAN) for a full-image enhancement. The model is available via the official GPEN
This specific model file represents one of the most powerful tools available for turning blurry, pixelated, or degraded faces into crystal-clear, high-resolution portraits. : Revitalizing blurry or grainy family historical photos
: Revitalizing blurry or grainy family historical photos into sharp, modern resolutions.
| Dataset | Size | Content | |---------|------|---------| | (official StyleGAN2 pre‑training) | 70 k high‑quality portraits | Balanced gender/ethnicity, diverse ages, backgrounds. | | Synthetic Degradation Pipeline (used for BFR) | N/A (on‑the‑fly) | Randomly sampled combinations of: • Down‑sampling factors (2‑× to 16‑×) • Gaussian blur (σ = 0‑3) • Motion blur (kernel lengths up to 25 px) • JPEG compression (Q = 10‑100) • Additive Gaussian noise (σ = 0‑25) • Random color shift (γ, contrast). | | Real‑World BFR Test Set (e.g., CelebA‑HQ degraded, LFW‑BFR) | 5 k images | For evaluation only, not used in training. |
You should consider using gpen-bfr-2048.pth if: