Trained a multiClass classification Convolution neural network by TransferLearning using MobileNetV2 with imagenet weights.Which takes image data and predicts the name of fruits and vegetables.Trained on GoogleColab.
- Trained on GoogleColab.
- Image Data preprocessing.
- Data Visualization and Exploratory Data Analysis.
- Image Data Normalization and Scaling.
- Used image_dataset_from_directory for batch creation.
- Created augmented data configuration inside Sequential model layer.
- Trained the model in FunctionalApi config by transferLearning using MobileNetV2 with imagenet weights as base model.
- Also added multiple GlobalAveragePooling2D layers,Flatten and Dense layers on top of based model for output.
- Retrained the model after hyperparameter tuning with baselayers of pretrained model as trainable and acquired better accuracy.
- Tested the model on custom data.
https://www.kaggle.com/kritikseth/fruit-and-vegetable-image-recognition
- Tensorflow 2
- Keras
- Numpy
- Matlplotlib
- Seaborn
- os
- PIL
