#!/usr/bin/env python
# coding: utf-8

# In[1]:


import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns


# In[2]:


import tensorflow as tf
from tensorflow import keras
from keras import Sequential
from keras.layers import Dense,Convolution2D,Flatten,Dropout,BatchNormalization
from tensorflow.keras.layers import MaxPooling2D
from keras.preprocessing.image import ImageDataGenerator


# In[ ]:


#val_data=keras.utils.image_dataset_from_directory(
#directory="E:\DSspec\Internship\CUB-200-2011\cub_200_2011_64x64_for_fid_10k\cub_200_2011_64x64_10k"
#label="inferred",
#label_mode="int",
#batch_size=32,
#iamge_size=(256,256)
#)


# In[3]:


train=keras.utils.image_dataset_from_directory(directory="E:\\DSspec\\Internship\\CUB-200-2011\\cub_200_2011_64x64_for_fid_10k",
                                             labels="inferred",
                                             validation_split=0.2,
                                             subset="training",
                                             seed=1337,
                                             label_mode="int",
                                             batch_size=32,
                                             image_size=(256,256))


# In[4]:


test=keras.utils.image_dataset_from_directory(directory="E:\\DSspec\\Internship\\CUB-200-2011\\cub_200_2011_64x64_for_fid_10k",
                                             labels="inferred",
                                             validation_split=0.2,
                                             subset="validation",
                                             seed=1337,
                                             label_mode="int",
                                             batch_size=32,
                                             image_size=(256,256))


# In[5]:


for image,label in train.take(2):
    plt.imshow(image[31].numpy().astype("uint8"))
    plt.show()


# In[8]:


from tensorflow.keras import layers 
data_augmentation = keras.Sequential(
  [
    layers.RandomFlip("horizontal", input_shape=(256, 256, 3)),
    layers.RandomRotation(0.3),
    layers.RandomZoom(0.3),
  ]
)


# In[9]:


train_gen = train.map(lambda x, y: (data_augmentation(x, training=True), y))


# In[ ]: