LaMa (ONNX, 512 × 512)

LaMa (big-lama) for image inpainting, as ONNX with a fixed 512 × 512 input. Its Fourier convolutions are ONNX DFT ops. Card Anvil uses it in the browser to remove watermarks from card art.

This is an unchanged copy of one file from Carve/LaMa-ONNX at commit c3c0c9e, kept here so the app loads a file that can't change under it.

File Source path
lama_fp32.onnx lama_fp32.onnx

Credits and licence

  • Model: LaMa by Samsung AI Center, from Resolution-robust Large Mask Inpainting with Fourier Convolutions (WACV 2022).
  • ONNX port: Carve (Carve/LaMa-ONNX).

Licensed under the Apache License 2.0, as the source is. See LICENSE.

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support