|
|
--- |
|
|
license: mit |
|
|
datasets: |
|
|
- cifar100 |
|
|
language: |
|
|
- en |
|
|
pipeline_tag: image-classification |
|
|
--- |
|
|
# micromind checkpoints for CIFAR-100 |
|
|
|
|
|
This repository contains checkpoints for the CIFAR-100 dataset for the following networks: |
|
|
|
|
|
| Model | Top 1 Accuracy | Top 5 Accuracy | |
|
|
| ------------------ |---------------- | -------------- | |
|
|
| `PhiNet(alpha=3, beta=0.75, t_zero=6, num_layers=7, resolution=160)` | 75.56% | 93.5% | |
|
|
| `PhiNet(alpha=0.75, beta=1, t_zero=6, num_layers=5, resolution=160)` | 60.87% | 86.98% | |
|
|
| `PhiNet(alpha=0.35, beta=1, t_zero=6, num_layers=7, resolution=160)` | 59.77% | 85.98% | |
|
|
| `PhiNet(alpha=0.25, beta=1, t_zero=6, num_layers=7, resolution=160)` | 54.16% | 82.32% | |
|
|
| `PhiNet(alpha=0.25, beta=1, t_zero=5, num_layers=7, resolution=160)` | 52.41% | 81.24% | |
|
|
|
|
|
To download and use this repo: |
|
|
|
|
|
``` |
|
|
from micromind import PhiNet |
|
|
model = PhiNet.from_pretrained("CIFAR-100", alpha=3.0, beta=0.75, t_zero=6, num_layers=7, num_classes=100, resolution=160) |
|
|
``` |
|
|
|
|
|
## Authors |
|
|
|
|
|
- [@fpaissan](https://www.github.com/fpaissan) |
|
|
- [@matteobeltrami](https://www.github.com/matteobeltrami) |
|
|
|
|
|
--- |
|
|
license: mit |
|
|
--- |
