You can deploy this model and start earning money today!

vrudish-waghmare/
food-vision-big-model

Name: food-vision-big-model
Availability: InStock

Views

No views yet

🍔👁 Food Vision Big – Food Image Classification Model

This model is a deep learning-based image classifier trained to identify food items across 101 categories using the Food101 dataset.

It leverages transfer learning with EfficientNet and is fine-tuned using mixed precision training for improved performance and efficiency.

Architecture: EfficientNet (Transfer Learning)
Framework: TensorFlow / Keras
Input Size: 224 × 224 × 3
Output: 101 food classes
Training Strategy:
- Feature Extraction (frozen base model)
- Fine-Tuning (all layers unfrozen)
Learning Rate:
- Initial: 1e-3
- Fine-Tuning: 1e-4
Precision: Mixed Precision (float16)

Images are:

The model achieves the following performance on the test dataset:

This demonstrates strong generalization across a large and diverse food dataset.

python

import tensorflow as tf
import tf_keras as keras

model = keras.models.load_model("food_vision_big_model.keras")

⚠️ Limitations

📌 Intended Use

👨‍💻 Author

Vrudish Waghmare
Machine Learning & AI Enthusiast