metadata
language: en
license: mit
datasets:
- Cats-Dogs
metrics:
- accuracy
- f1
- precision
- recall
Transfer Learning On Alexnet:
This model uses Transfer Learning for a 2 labelled class Cats-Dogs classification.
- Pre-trained model used: Alexnet.
- Dataset used: Cats_Dogs.
- batch_size = 8
- learning_rate = 0.001
- epochs = 5
Evaluation Results
📌 Before Transfer Learning
** Accuracy:** 40.66%
Precision: 0.3983,
Recall: 0.4066,
F1-score: 0.3942,
Confusion Matrix: [[ 659 1841] [1126 1374]]
✅ After Training
** Train Accuracy:** 92.40%,
- Precision: 0.9250,
- Recall: 0.9240,
- F1-score: 0.9240,
- Confusion Matrix: [[9486 514] [1006 8994]]
** Validatio Accuracy:** 94.10%
Precision: 0.9425,
Recall: 0.9410,
F1-score: 0.9410,
Confusion Matrix: [[2425 75] [ 220 2280]]
You can download the model from Hugging Face.