Decision Workspace
aegir vs torsh-autograd vs tang
Side-by-side comparison of Rust crates
44
aegir
growingv2.0.0
Strongly-typed, reverse-mode autodiff library in Rust
62
torsh-autograd
growingv0.2.0
Automatic differentiation engine for ToRSh with PyTorch-compatible API
58
tang
experimentalv0.2.1
Math library for physical reality — geometry, spatial algebra, tensor, training, GPU compute, and 3D gaussian splatting
Core Metrics
| aegir | torsh-autograd | tang | |
|---|---|---|---|
| Health Score | 44 | 62 | 58 |
| Total Downloads | 4.7K | 3.2K | 1.6K |
| 30d Downloads | 0 | 0 | 0 |
| Dependents | 0 | 57 | 28 |
| Releases | 3 | 9 | 3 |
| Last Updated | 1380d ago | 57d ago | 30d ago |
| Age | 5y 9m | 1y | 7m |
Health Breakdown
aegir
Maintenance
4
Quality
19
Community
3
Popularity
4
Documentation
14
torsh-autograd
Maintenance
19
Quality
13
Community
11
Popularity
4
Documentation
15
tang
Maintenance
17
Quality
15
Community
10
Popularity
4
Documentation
12
Technical Details
| aegir | torsh-autograd | tang | |
|---|---|---|---|
| Version | 2.0.0 | 0.2.0 | 0.2.1 |
| Stable (≥1.0) | ✓ Yes | ✗ No | ✗ No |
| License | MIT | Apache-2.0 | MIT |
| Dependencies | 10 | 23 | 15 |
| Crate Size | 46KB | 829KB | 285KB |
| Features | 1 | 9 | 22 |
| Yanked % | 0.0% | 0.0% | 0.0% |
| Edition | 2018 | 2021 | 2021 |
| MSRV | 1.66 | 1.87 | 1.94 |
| Owners | 1 | 1 | 1 |
Links
Quick Verdict
- •torsh-autograd leads with a health score of 62/100, but none of the options score above 80.
- •aegir has the most downloads (4.7K), suggesting wider adoption.
- •⚠ aegir has not been updated in over a year.
- •torsh-autograd, tang are pre-1.0 — API may change.