sheldonxxxx/RapidRAW-Nonlocal-Denoise

RapidRAW Nonlocal RAW Denoise — runtime artifacts

This repository distributes runtime conversion artifacts for the RapidRAW Nonlocal Bayer RAW denoiser. It is not a newly trained model: both artifacts are converted, without retraining or precision changes, from the same pinned upstream checkpoint.

Upstream attribution and license

• Project: MIA-UIB/nonlocal-matchfilter
• Paper: “Learned Nonlocal Feature Matching and Filtering for RAW Image Denoising”, Marco Sánchez-Beeckman and Antoni Buades, arXiv:2604.17453.
• Source checkpoint: upstream v1.0.0 release weights (rawnoise_25_15_9nbr.ckpt), tensor-only SHA-256 c16747d852b93a95908792cdbac901f89cca98e35b91ea7db6214de42fbd3cad.
• License: MIT (see LICENSE; copyright Marco Sánchez Beeckman). Attribution retained per the license terms.

Input contract (both variants)

• RAW Bayer input packed to 4 RGBG planes plus a 4-channel noise-level map: raw_with_noise [1, 8, 320, 320], float32.
• Output: denoised packed RAW denoised_raw [1, 4, 320, 320], float32.
• Tiling: 320 px packed tiles with 64 px halo (192 px retained core).
• Host ensemble: e1 (one pass, balanced) or e4 (four rotations, maximum).

Variants

Variant	Backend	File	SHA-256
CoreML (macOS, direct CoreML.framework)	native-coreml-v1	coreml/packed.mlpackage	ecf8f41b72b8e79a7210f61b13b3a21ad2ddbc79c475039be367fc011a98f5d3
ONNX (CPU / Linux CUDA)	native-onnx-v1	onnx/model.onnx	df7f21ddbdfecd0984896a902623f75aa883d25b5f862c49d8c2b3f63a7dd2d9

• CoreML is a direct TorchScript conversion (coremltools-direct-torchscript, FLOAT32, static-packed sampler) for native CoreML.framework inference on macOS. It is not an ONNX/CoreML-execution-provider artifact.
• distribution.json lists every downloadable file with size and SHA-256. RapidRAW downloads by immutable commit revision and verifies distribution.json, every file, and the inner bundle contracts — never main alone.

Measured evidence (honest scope)

• Historical study (Python reference tooling): 18/18 fixture tiles PASS and six full-photo runs PASS under frozen gates.
• Current-source native integration: 18/18 fixture tiles PASS through the Rust CoreML backend, plus portrait full-photo e1 and e4 PASS. Broader current-source 6/6 full-photo reruns were not performed.
• Experimental: commercial parity has not been established.

Intended use

• Automatic installation by RapidRAW (install_model kind=nonlocal).
• No hosted inference API is provided or expected.