Gradient and training anomaly monitoring for PyTorch and TensorFlow
Project description
GradScope
GradScope is a small, focused library that watches your training loops and tells you when gradients, metrics, or weight updates go bad.
- 2-line attach for PyTorch and TensorFlow
- Automatic collection of gradients, updates, drift, and metrics
- Rich alerting, summaries, diffing, export, CLI, and web API
- Lightweight FastAPI dashboard for quick inspection of runs
- Git-aware run metadata (commit, branch, dirty state) for reproducibility
For full documentation and examples, see docs/README.md.
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
gradscope-0.1.0.tar.gz
(20.8 kB
view details)
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
gradscope-0.1.0-py3-none-any.whl
(17.9 kB
view details)
File details
Details for the file gradscope-0.1.0.tar.gz.
File metadata
- Download URL: gradscope-0.1.0.tar.gz
- Upload date:
- Size: 20.8 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.13.3
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
f1038813b4074a80c06cb18a8a36142f13c087c25c424366f30948b13012383d
|
|
| MD5 |
77c0409fe841f2146a7a9e81dd705dc7
|
|
| BLAKE2b-256 |
a81216b01e7f7fb93a4a2385a3469e50cdedad2a2a77fe416cd92dacccf24c4f
|
File details
Details for the file gradscope-0.1.0-py3-none-any.whl.
File metadata
- Download URL: gradscope-0.1.0-py3-none-any.whl
- Upload date:
- Size: 17.9 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.2.0 CPython/3.13.3
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
d1450af544c87f4c4b6e0993cd230f32793a7f184e54f33b2da697a5224adf0b
|
|
| MD5 |
f9137262be3bcb6c2bac2fdaa1eb51a9
|
|
| BLAKE2b-256 |
65f09935a4e82af87612c232261b7dd7deb29d3f8886291b16ecf59dd93aa84d
|
