Created
September 9, 2019 21:54
-
-
Save rezapci/0d4fb510aa55f58a7b960e255658e150 to your computer and use it in GitHub Desktop.
neural_network.py.ipynb
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
{ | |
"cells": [ | |
{ | |
"metadata": { | |
"ExecuteTime": { | |
"start_time": "2019-09-09T21:54:32.444Z" | |
}, | |
"trusted": true | |
}, | |
"cell_type": "code", | |
"source": "import numpy as np\nimport tensorflow as tf\nfrom utils import get_mnist_data\n\n\ndef single_layer_net(x):\n\n input_dim = 784\n n_classes = 10\n\n with tf.variable_scope('single_layer_net'):\n\n W = tf.Variable(initial_value=tf.random_normal(shape=[input_dim, n_classes]), name='weights')\n b = tf.Variable(initial_value=tf.zeros(shape=[n_classes]), name='biases')\n\n logits = tf.matmul(x, W) + b\n\n y = tf.nn.softmax(logits)\n\n return y\n\n\ndef multi_layer_net(x):\n\n input_dim = 784\n hidden_dim = 100\n n_classes = 10\n\n with tf.variable_scope('multi_layer_net'):\n\n W_1 = tf.Variable(initial_value=tf.random_normal(shape=[input_dim, hidden_dim]), name='l1_weights')\n b_1 = tf.Variable(initial_value=tf.zeros(shape=[hidden_dim]), name='l1_biases')\n\n hidden_1 = tf.nn.relu(tf.matmul(x, W_1) + b_1)\n\n W_2 = tf.Variable(initial_value=tf.random_normal(shape=[hidden_dim, n_classes]), name='l2_weights')\n b_2 = tf.Variable(initial_value=tf.zeros(shape=[n_classes]), name='l2_biases')\n\n logits = tf.matmul(hidden_1, W_2) + b_2\n\n y = tf.nn.softmax(logits)\n\n return y\n\n\nif __name__ == '__main__':\n\n # Load MNIST data\n mnist = get_mnist_data('/tmp/mnist', verbose=True)\n\n # Placeholders\n x = tf.placeholder(dtype=tf.float32, shape=[None, 784]) # input placeholder\n\n # Placeholder for targets\n targets = tf.placeholder(dtype=tf.float32, shape=[None, 10])\n\n # Define model output\n y = multi_layer_net(x)\n\n # Define loss function\n loss = tf.reduce_mean(-tf.reduce_sum(targets * tf.log(y + np.finfo('float32').eps), axis=1))\n\n # Define train step\n train_step = tf.train.AdamOptimizer(learning_rate=0.001).minimize(loss)\n\n init_op = tf.global_variables_initializer()\n\n # Define metrics\n correct_predictions = tf.equal(tf.argmax(y, axis=1), tf.argmax(targets, axis=1))\n accuracy = tf.reduce_mean(tf.cast(correct_predictions, tf.float32))\n\n with tf.Session() as sess:\n\n # Initialize all variables\n sess.run(init_op)\n\n # Training parameters\n training_epochs = 100\n batch_size = 128\n\n # Number of batches to process to see whole dataset\n batches_each_epoch = mnist.train.num_examples // batch_size\n\n for epoch in range(training_epochs):\n\n # During training measure accuracy on validation set to have an idea of what's happening\n val_accuracy = sess.run(fetches=accuracy,\n feed_dict={x: mnist.validation.images, targets: mnist.validation.labels})\n print('Epoch: {:06d} - VAL accuracy: {:.03f}'.format(epoch, val_accuracy))\n\n for _ in range(batches_each_epoch):\n\n # Load a batch of training data\n x_batch, target_batch = mnist.train.next_batch(batch_size)\n\n # Actually run one training step here\n sess.run(fetches=[train_step],\n feed_dict={x: x_batch, targets: target_batch})\n\n # Eventually evaluate on whole test set when training ends\n test_accuracy = sess.run(fetches=accuracy,\n feed_dict={x: mnist.test.images, targets: mnist.test.labels})\n print('*' * 50)\n print('Training ended. TEST accuracy: {:.03f}'.format(test_accuracy))", | |
"execution_count": null, | |
"outputs": [] | |
}, | |
{ | |
"metadata": { | |
"trusted": true | |
}, | |
"cell_type": "code", | |
"source": "", | |
"execution_count": null, | |
"outputs": [] | |
} | |
], | |
"metadata": { | |
"kernelspec": { | |
"name": "python3", | |
"display_name": "Python 3", | |
"language": "python" | |
}, | |
"language_info": { | |
"codemirror_mode": { | |
"name": "ipython", | |
"version": 3 | |
}, | |
"file_extension": ".py", | |
"mimetype": "text/x-python", | |
"name": "python", | |
"nbconvert_exporter": "python", | |
"pygments_lexer": "ipython3", | |
"version": "3.7.3" | |
}, | |
"latex_envs": { | |
"eqNumInitial": 1, | |
"eqLabelWithNumbers": true, | |
"current_citInitial": 1, | |
"cite_by": "apalike", | |
"bibliofile": "biblio.bib", | |
"LaTeX_envs_menu_present": true, | |
"labels_anchors": false, | |
"latex_user_defs": false, | |
"user_envs_cfg": false, | |
"report_style_numbering": false, | |
"autoclose": false, | |
"autocomplete": true, | |
"hotkeys": { | |
"equation": "Ctrl-E", | |
"itemize": "Ctrl-I" | |
} | |
}, | |
"nbTranslate": { | |
"hotkey": "alt-t", | |
"sourceLang": "en", | |
"targetLang": "fr", | |
"displayLangs": [ | |
"*" | |
], | |
"langInMainMenu": true, | |
"useGoogleTranslate": true | |
}, | |
"toc": { | |
"nav_menu": {}, | |
"number_sections": true, | |
"sideBar": true, | |
"skip_h1_title": false, | |
"base_numbering": 1, | |
"title_cell": "Table of Contents", | |
"title_sidebar": "Contents", | |
"toc_cell": false, | |
"toc_position": {}, | |
"toc_section_display": true, | |
"toc_window_display": false | |
}, | |
"gist": { | |
"id": "", | |
"data": { | |
"description": "neural_network.py.ipynb", | |
"public": true | |
} | |
} | |
}, | |
"nbformat": 4, | |
"nbformat_minor": 2 | |
} |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment