DigiNews

Tech Watch by Johan Denoyer

← Back to articles

Show HN: Classify mechanical faults using Contrastive Language-Audio Pretraining

Quality: 8/10 Relevance: 9/10

Summary

This Show HN presents an open-source end-to-end audio-ML pipeline to diagnose car faults from noisy clips. It cleans audio, uses CLAP embeddings, and trains lightweight heads to triage faults with calibrated uncertainty; includes quickstart, demos, and docs.

🚀 Service construit par Johan Denoyer