|
--- |
|
license: mit |
|
datasets: |
|
- garythung/trashnet |
|
language: |
|
- en |
|
metrics: |
|
- accuracy |
|
- confusion_matrix |
|
- f1 |
|
--- |
|
|
|
# Trash Classifier Model |
|
|
|
This model is a fine-tuned ResNet50 for classifying trash images into categories such as plastic, paper, glass, metal, and others. |
|
|
|
## Dataset |
|
- Dataset: [TrashNet](https://huggingface.co/datasets/garythung/trashnet) |
|
- Number of Classes: 6 (Plastic, Metal, Paper, Glass, Cardboard, Trash) |
|
|
|
## Model Details |
|
- **Base Model**: ResNet50 (pretrained on ImageNet) |
|
- **Fine-tuning**: Input layer adjusted for 6 classes. |
|
- **Input Shape**: (384, 384, 3) |
|
- **Learning Rate**: 0.0001 |
|
- **Optimizer**: Adam |
|
|
|
## Performance |
|
- **Accuracy F1-Scores**: 92% |
|
|
|
## How to Use |
|
Load the model and use it for inference: |
|
```python |
|
import tensorflow as tf |
|
model = tf.keras.models.load_model("Stefaron/trash-classifier/best_model.keras") |
|
|
