Decision Workspace
rusty-gradients vs burn-cuda vs ghostflow-cuda
Side-by-side comparison of Rust crates
50
rusty-gradients
experimentalv0.2.0
A full-stack deep learning framework in Rust for training and deploying Transformer models. Features multi-backend support (CPU/CUDA/Metal/WASM), 62x GPU acceleration, Safetensors serialization, and BPE tokenization.
66
burn-cuda
growingv0.20.1
CUDA backend for the Burn framework
54
ghostflow-cuda
experimentalv1.0.0
CUDA backend for GhostFlow ML framework
Core Metrics
| rusty-gradients | burn-cuda | ghostflow-cuda | |
|---|---|---|---|
| Health Score | 50 | 66 | 54 |
| Total Downloads | 683 | 436.4K | 221 |
| 30d Downloads | 5 | 56.6K | 26 |
| Dependents | 0 | 131 | 8 |
| Releases | 3 | 19 | 2 |
| Last Updated | 67d ago | 25d ago | 78d ago |
| Age | 7m | 1y 7m | 2m |
Health Breakdown
rusty-gradients
Maintenance
15
Quality
13
Community
6
Popularity
3
Documentation
13
burn-cuda
Maintenance
18
Quality
17
Community
11
Popularity
7
Documentation
13
ghostflow-cuda
Maintenance
10
Quality
19
Community
9
Popularity
3
Documentation
13
Technical Details
| rusty-gradients | burn-cuda | ghostflow-cuda | |
|---|---|---|---|
| Version | 0.2.0 | 0.20.1 | 1.0.0 |
| Stable (≥1.0) | ✗ No | ✗ No | ✓ Yes |
| License | MIT | MIT OR Apache-2.0 | MIT OR Apache-2.0 |
| Dependencies | 26 | 4 | 3 |
| Crate Size | 129KB | 29KB | 29KB |
| Features | 15 | 7 | 2 |
| Yanked % | 0.0% | 0.0% | 0.0% |
| Edition | 2021 | 2024 | 2021 |
| MSRV | — | — | — |
| Owners | 1 | 1 | 1 |
Links
Quick Verdict
- •burn-cuda leads with a health score of 66/100, but none of the options score above 80.
- •burn-cuda is depended on by 131 crates — strongest ecosystem trust.
- •rusty-gradients, burn-cuda are pre-1.0 — API may change.