resnet18_cifar100 / README.md
edadaltocg's picture
Upload README.md with huggingface_hub
2964102
|
raw
history blame
1.24 kB
---
language: en
license: mit
library_name: timm
tags:
- image-classification
- resnet18
- cifar100
datasets: cifar100
metrics:
- accuracy
model-index:
- name: resnet18_cifar100
results:
- task:
type: image-classification
dataset:
name: CIFAR-100
type: cifar100
metrics:
- type: accuracy
value: 0.7843
---
# Model Card for Model ID
This model is a small resnet18 trained on cifar100.
- **Test Accuracy:** 0.7843
- **License:** MIT
## How to Get Started with the Model
Use the code below to get started with the model.
```python
import detectors
import timm
model = timm.create_model("resnet18_cifar100", pretrained=True)
```
## Training Data
Training data is cifar100.
## Training Hyperparameters
- config: scripts/train_configs/cifar100.json
- model: resnet18_cifar100
- dataset: cifar100
- batch_size: 64
- epochs: 200
- validation_frequency: 5
- seed: 1
- criterion: CrossEntropyLoss
- criterion_kwargs: {}
- optimizer: SGD
- lr: 0.1
- optimizer_kwargs: {'momentum': 0.9, 'weight_decay': 0.0005}
- scheduler: CosineAnnealingLR
- scheduler_kwargs: {'T_max': 190}
- debug: False
## Testing Data
Testing data is cifar100.
This model card was created by Eduardo Dadalto.