ModelAutopsy: High-Performance ML Failure Detection
Project description
ModelAutopsy
ModelAutopsy is a high-performance, hybrid-engine (Rust & C++) Machine Learning debugging library. It acts as a real-time diagnostic tool for Deep Learning pipelines, instantly detecting NaN, Infinity, Vanishing, and Exploding gradients.
Features
- Rust Safety Core: Uses Rust's memory safety guarantees to prevent Python crashes (Segfaults) during inspection.
- C++ Speed Core: Uses OpenMP SIMD instructions to analyze 1M+ parameters in milliseconds.
- Zero-Overhead: Uses direct memory access to NumPy buffers, avoiding Python GIL.
- IDE-Style Debugging: Includes a
@watchdecorator that acts like a live breakpoint for your training loops. - Advanced Metrics: Calculates Variance, L2 Norm, Mean, Min/Max in a single pass.
Installation
pip install modelautopsy
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
modelautopsy-1.0.5.tar.gz
(7.1 kB
view details)
File details
Details for the file modelautopsy-1.0.5.tar.gz.
File metadata
- Download URL: modelautopsy-1.0.5.tar.gz
- Upload date:
- Size: 7.1 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.14.2
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
1862d20131da065543d9e105ba1a3404854eacd5097ba95ea4be6d7c1b950983
|
|
| MD5 |
e218051be696ec04d1c2e6fe05ca9d0d
|
|
| BLAKE2b-256 |
cbd2927518123ab6a490a88bc865b8fb5b8e68320a0ac43d3ff62c09f6bcd762
|
