adobe-research/convmelspec
Convmelspec: Convertible Melspectrograms via 1D Convolutions
This project helps machine learning engineers and researchers deploy audio-based AI models to mobile devices or other edge environments. It converts audio features, specifically Mel-spectrograms, into a format compatible with on-device machine learning frameworks like CoreML and ONNX. You input a trained audio model (e.g., in PyTorch), and it outputs a portable model file ready for deployment, enabling your audio AI to run efficiently on various platforms.
147 stars. No commits in the last 6 months.
Use this if you need to deploy your audio machine learning model to a mobile device or embedded system and require a Mel-spectrogram computation to be part of the on-device inference.
Not ideal if you are only training audio models and do not need to deploy them to a cross-platform, on-device environment, or if your model does not rely on Mel-spectrogram features.
Stars
147
Forks
10
Language
Python
License
Apache-2.0
Category
Last pushed
May 13, 2024
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/adobe-research/convmelspec"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
apple/coremltools
Core ML tools contain supporting tools for Core ML model conversion, editing, and validation.
hollance/YOLO-CoreML-MPSNNGraph
Tiny YOLO for iOS implemented using CoreML but also using the new MPS graph API.
tensorflow/swift-apis
Swift for TensorFlow Deep Learning Library
hanleyweng/CoreML-in-ARKit
Simple project to detect objects and display 3D labels above them in AR. This serves as a basic...
hollance/Forge
A neural network toolkit for Metal