Decision Workspace
float4 vs aprender-train vs torsh-quantization
Side-by-side comparison of Rust crates
55
float4
growingv0.2.0
MXFP4-compatible 4-bit floating point types and block formats for Rust.
61
aprender-train
experimentalv0.70.3
Training & Optimization library with autograd, LoRA, quantization, and model merging
60
torsh-quantization
growingv0.2.0
Model quantization for ToRSh neural networks
Core Metrics
| float4 | aprender-train | torsh-quantization | |
|---|---|---|---|
| Health Score | 55 | 61 | 60 |
| Total Downloads | 1.2M | 6.3K | 822 |
| 30d Downloads | 0 | 0 | 0 |
| Dependents | 56 | 253 | 10 |
| Releases | 2 | 31 | 9 |
| Last Updated | 219d ago | 1d ago | 57d ago |
| Age | 1y 2m | 6m | 1y |
Health Breakdown
float4
Maintenance
8
Quality
16
Community
11
Popularity
7
Documentation
13
aprender-train
Maintenance
20
Quality
11
Community
13
Popularity
5
Documentation
12
torsh-quantization
Maintenance
19
Quality
14
Community
9
Popularity
3
Documentation
15
Technical Details
| float4 | aprender-train | torsh-quantization | |
|---|---|---|---|
| Version | 0.2.0 | 0.70.3 | 0.2.0 |
| Stable (≥1.0) | ✗ No | ✗ No | ✗ No |
| License | MIT | MIT | Apache-2.0 |
| Dependencies | 0 | 66 | 11 |
| Crate Size | 17KB | 1.9MB | 288KB |
| Features | 0 | 19 | 3 |
| Yanked % | 0.0% | 3.2% | 0.0% |
| Edition | 2024 | 2021 | 2021 |
| MSRV | — | 1.87 | 1.87 |
| Owners | 1 | 1 | 1 |
Links
Quick Verdict
- •aprender-train leads with a health score of 61/100, but none of the options score above 80.
- •float4 has the most downloads (1.2M), suggesting wider adoption.
- •aprender-train is depended on by 253 crates — strongest ecosystem trust.