kaggle-Flower-Classification-on-TPU

This Project is based on Kaggle's Competition : Flower classification on TPU

About Challenge

Classify 104 different types of flowers

Data Description

Images are provided in TFRecord format, a container format frequently used in Tensorflow to group data files for optimal training performace. Each file contains the id, label and image.

12753 training images
3712 validation images
7382 unlabeled test images

Data is provided for resolutions:

(512 X 512 X 3)
(331 X 331 X 3)
(224 X 224 X 3)
(192 X 192 X 3)

Pipeline:

Refer to Kaggle Notebook : here for the approach and implementation.

Results

Model	Score
MobilenetV2	0.93745
DenseNet201	0.96971
EfficientNetB7	0.95150
(0.62)DenseNet201 + (1-0.62)EfficientNetB7	0.97058

Strategy

Input Data Pipeline:
- tf.data pipeline with data augmentation
- Added additional data
- Data Augmentation done are:
  1. Spatial level Transforms:
    - Random flip left right, Random flip up down, Random crop
  2. Pixel level Transforms :
    - Random Saturation, Random Contrast, Adjust Gamma, Random Brightness
  3. Miscellaneous :
    - Transform Rotation, Transform Shift, Transform Shear, Transform Zoom
Model
- BaseModel with input image size of 224 X 224 and Added GlobalAveragePooling layer
- Transfer learning over several models out which DenseNet201 performed best
- Ensemble Learning with 2 or more models
  - used DenseNet201 and EfficientnetB7 :
    - instead of one network, we train two and then combine their probability distributions.
Training parameters
- Hyper-parameters : Image Size, Batch Size, Epochs, Learning rate, Types and Amount of Augmentation
- used learning rate scheduling, for more stable training.
  - High initial learning rates can make loss explode. It is usually better linearly to increase learning rate from very small value over the first ~5 iterations.
```
LR_START = 0.00001
LR_MAX = 0.00005 * strategy.num_replicas_in_sync
```
- optimizer = Adam
- loss function = sparse_categorical_crossentropy
- metrics = sparse_categorical_accuracy

Predictions on real images

marigold	rose	sunflower	dahlia

Extended my Work to build Android application for Flower classification

Converted trained model to TFLite Format
Did setup for Android App in Android studio
apk files are present here

ankita-2015/kaggle-Flower-Classification-on-TPU