Image eraser (browser build)
Unmodified inpainting_lama_2025jan.onnx (92,591,623 bytes) from opencv/inpainting_lama revision aee6d22f0a13e5e35af1c9a1c3afd62841fc6f3f.
Apache License 2.0 (see LICENSE). Upstream model card: LaMa (advimman/lama), the quantized ONNX published by OpenCV.
Used the same way as WaveSpeed's image eraser: inputs image float32 [1,3,512,512] BGR scaled by 1/255, and mask float32 [1,1,512,512] with 1 on pixels to fill. Output output float32 [1,3,512,512] in 0–255 BGR.
Xet Storage Details
- Size:
- 597 Bytes
- Xet hash:
- c066a7863eb0c6501aa9b78cddf44fd751eb5bf396bbbada75c3ce3a84e6ed13
·
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.