AI article

FP16 Quantization Made Our In-Browser Real-ESRGAN 4.8x Faster — and Nearly Destroyed Our Inpainting Model

Community description: We run super-resolution and watermark removal entirely in the browser at ClearPix, using ONNX Runtime...

Dev.to | Oct 4, 2026 | Umasou

Automated excerpt

The WASM EP also runs the fp16 model fine, so our Chromium e2e path needed no special-casing. Our watermark removal pipeline uses LaMa, an inpainting model. Pick a model designed for the deployment environment.

Selected automatically from source text; not independently written or fact-checked. Read the original for full context.

Read the original article

AI briefing: recent picks

More AI news