Decision Workspace
float4 vs torsh-quantization vs aprender-train
Side-by-side comparison of Rust crates
56
float4
experimentalv0.2.0
MXFP4-compatible 4-bit floating point types and block formats for Rust.
61
torsh-quantization
experimentalv0.1.3
Model quantization for ToRSh neural networks
57
aprender-train
experimentalv0.60.0
Training & Optimization library with autograd, LoRA, quantization, and model merging
Core Metrics
| float4 | torsh-quantization | aprender-train | |
|---|---|---|---|
| Health Score | 56 | 61 | 57 |
| Total Downloads | 535.1K | 720 | 3.8K |
| 30d Downloads | 65.2K | 0 | 0 |
| Dependents | 39 | 9 | 150 |
| Releases | 2 | 8 | 18 |
| Last Updated | 131d ago | 13d ago | 7d ago |
| Age | 11m | 9m | 3m |
Health Breakdown
float4
Maintenance
11
Quality
15
Community
10
Popularity
7
Documentation
13
torsh-quantization
Maintenance
21
Quality
13
Community
9
Popularity
3
Documentation
15
aprender-train
Maintenance
20
Quality
9
Community
12
Popularity
4
Documentation
12
Technical Details
| float4 | torsh-quantization | aprender-train | |
|---|---|---|---|
| Version | 0.2.0 | 0.1.3 | 0.60.0 |
| Stable (≥1.0) | ✗ No | ✗ No | ✗ No |
| License | MIT | Apache-2.0 | MIT |
| Dependencies | 0 | 11 | 66 |
| Crate Size | 17KB | 280KB | 2.0MB |
| Features | 0 | 3 | 19 |
| Yanked % | 0.0% | 0.0% | 5.6% |
| Edition | 2024 | 2021 | 2021 |
| MSRV | — | 1.77 | 1.87 |
| Owners | 1 | 1 | 1 |
Links
Quick Verdict
- •torsh-quantization leads with a health score of 61/100, but none of the options score above 80.
- •float4 has the most downloads (535.1K), suggesting wider adoption.
- •aprender-train is depended on by 150 crates — strongest ecosystem trust.