mlboydaisuke/Real-ESRGAN-x4-CoreML

Real-ESRGAN 4x — Core ML

Super Resolution, 2021

Real-world blind super-resolution. 4× upscale from any input. Handles noise, blur, and JPEG artifacts. 512×512 input → 2048×2048 output.

Core ML conversion of xinntao/Real-ESRGAN for on-device inference on iPhone, iPad and Mac. Converted with coremltools; the packages are stateless, so all sequencing and buffering lives in your Swift code.

Task	image to image
Upstream	xinntao/Real-ESRGAN
Packages	1
Download size	59 MB
Minimum iOS	17.0
Peak RAM	~500 MB

Files

File	Size	Compute units	SHA-256
RealESRGAN_x4.mlpackage.zip	59 MB	all	5fd9d7ea7e6187ff…
Total	59 MB

compute_units is not a suggestion -- it is the configuration the conversion was verified against. Moving a package to a different compute unit can silently change the numerics (FP16 attention overflow) or crash on the GPU.

Download

hf download mlboydaisuke/coreml-zoo --include "realesrgan/*" --local-dir ./realesrgan
unzip './realesrgan/realesrgan/*.zip' -d ./realesrgan

Use in Swift

import CoreML

let config = MLModelConfiguration()
config.computeUnits = .all   // as converted — see the table above

// Unzip the .mlpackage, drop it into your Xcode target and Xcode compiles it
// at build time:
let model = try RealESRGAN_x4(configuration: config)

// ...or compile a downloaded .mlpackage at runtime:
let compiled = try await MLModel.compileModel(at: mlpackageURL)
let model = try MLModel(contentsOf: compiled, configuration: config)

Demo

• Models Zoo — this model is downloadable and runnable inside the Models Zoo app on the App Store, no build required.

Conversion

• Pitfalls hit during conversion (FP16 overflow, ANE buffer limits, stride handling): docs/coreml_conversion_notes.md
• Model index: CoreML-Models

License

The conversion inherits the upstream license: BSD-3-Clause.

Credits

• Upstream authors: xinntao/Real-ESRGAN, 2021
• Core ML conversion: john-rocky (Daisuke Majima)