Phips/nerve

NERVE — image upscaling & restoration

Drop an image, pick a model, press Upscale, and download the full-quality PNG.

What is NERVE? A small (~1.8M params) pure-convolution network built for ease of use, not benchmark chasing: one file, one variant, and a checkpoint that converts cleanly to dynamic ONNX, TensorRT and ncnn — no fused/unfused pairs.

• —2x / 4x release → for clean images.
• —2x / 4x OTF GAN → for degraded images (blurry, noisy, JPEG-compressed).

Default is the 2x OTF GAN (best all-round for degraded input); other models stay selectable.

Models, ONNX/ncnn files, training configs and the architecture live at [huggingface.co/Phips/NERVE](https://huggingface.co/Phips/NERVE).

NERVE links: models & configs · try it online (ZeroGPU) · local app · train your own (Colab)

Runs on ZeroGPU; large images are processed in tiles. Apache-2.0.