fynnsu

import json

import torch
from tokenizers import Tokenizer
from transformers import (
    AutoModelForCausalLM,
    AutoTokenizer,
    LlamaConfig,
    LlamaForCausalLM,
)

fynnsu

digit = 3 # The class to display average image of
plt.imshow(x_train[np.argmax(y_train, axis=1) == digit].mean(axis=0).reshape(28,28), cmap='gray')

fynnsu

index = 18
print(np.argmax(y_train[index]))
plt.imshow(x_train[index].reshape(28,28), cmap='gray')

fynnsu

BATCH_SIZE = 64
EPOCHS = 50
m = len(x_train)

sess = tf.Session()
sess.run(tf.global_variables_initializer())

fynnsu

for epoch in range(EPOCHS):
  for i in range(BATCH_SIZE, m, BATCH_SIZE):
      x_batch = x_train[i-BATCH_SIZE:i]
      y_batch = y_train[i-BATCH_SIZE:i]
      sess.run(step, feed_dict={X:x_batch, labels:y_batch})
  if (epoch+1) % 5 == 0:
      predictions, test_loss = sess.run([y, loss], feed_dict={X:x_test,  
          labels:y_test})
      accuracy = np.mean(np.argmax(y_test, axis=1) == np.argmax(predictions,  
          axis=1))

fynnsu

import tensorflow as tf
from matplotlib import pyplot as plt
import numpy as np

from tensorflow.examples.tutorials.mnist import input_data
mnist = input_data.read_data_sets('MNIST_data', one_hot=True)

fynnsu

# shuffle training data
combined = list(zip(x_train, y_train))
np.random.shuffle(combined)
x_train[:], y_train[:] = zip(*combined)

# shuffle test data
combined = list(zip(x_test, y_test))
np.random.shuffle(combined)
x_test[:], y_test[:] = zip(*combined)

fynnsu

print('x_train:', x_train.shape)
print('y_train:', y_train.shape)
print('x_test:', x_test.shape)
print('y_test:', y_test.shape)

fynnsu

predictions = sess.run(y, feed_dict={X:x_test})
index = 9
print('Prediction: %d, Label: %d' % (np.argmax(predictions[index]),  
  np.argmax(y_test[index])))
plt.imshow(x_test[index].reshape(28,28), cmap='gray')

fynnsu

loss = tf.losses.softmax_cross_entropy(onehot_labels=labels, logits=y)
step = tf.train.AdamOptimizer().minimize(loss)