cmc/data-science
Personal data science learning files.
clone: git clone https://gitbay.org/cmc/data-science.git
main: notebooks/TensorFlow_QuickStart.ipynb · raw
1{
2 "cells": [
3 {
4 "cell_type": "markdown",
5 "id": "368ae4ce-f2d4-4d31-b4b6-4c95c01c472c",
6 "metadata": {},
7 "source": [
8 "# TensorFlow Quickstart\n",
9 "\n",
10 "Getting started with neural network machine learning models in TensorFlow."
11 ]
12 },
13 {
14 "cell_type": "markdown",
15 "id": "f97fa6a5-b5db-45ac-98f4-54a4fee7ddaf",
16 "metadata": {},
17 "source": [
18 "## Set up TensorFlow"
19 ]
20 },
21 {
22 "cell_type": "code",
23 "execution_count": 3,
24 "id": "3ce17707-7c32-4ccf-8ef1-fbad5a78db7b",
25 "metadata": {},
26 "outputs": [],
27 "source": [
28 "# pip3 install tensorflow"
29 ]
30 },
31 {
32 "cell_type": "code",
33 "execution_count": 4,
34 "id": "9e0d2030-33c0-4da7-bf65-919cdb3c113c",
35 "metadata": {},
36 "outputs": [
37 {
38 "name": "stdout",
39 "output_type": "stream",
40 "text": [
41 "TensorFlow version: 2.13.0\n"
42 ]
43 }
44 ],
45 "source": [
46 "import tensorflow as tf\n",
47 "print(\"TensorFlow version:\", tf.__version__)"
48 ]
49 },
50 {
51 "cell_type": "markdown",
52 "id": "d23e7ecb-d531-426d-85dd-1d1d0f926439",
53 "metadata": {},
54 "source": [
55 "## Load a dataset"
56 ]
57 },
58 {
59 "cell_type": "code",
60 "execution_count": 6,
61 "id": "d04bb60f-346b-44f5-bae8-16bb1cc60f87",
62 "metadata": {},
63 "outputs": [],
64 "source": [
65 "# Load and prepare the MNIST dataset. The pixel values of the images range from 0 through 255.\n",
66 "# Scale these values to a range of 0 to 1 by dividing the values by 255.0.\n",
67 "# This also converts the sample data from integers to floating-point numbers:\n",
68 "mnist = tf.keras.datasets.mnist\n",
69 "\n",
70 "(x_train, y_train), (x_test, y_test) = mnist.load_data()\n",
71 "x_train, x_test = x_train / 255.0, x_test / 255.0"
72 ]
73 },
74 {
75 "cell_type": "code",
76 "execution_count": 19,
77 "id": "7895b2b4-3666-4a00-9536-8cdf4c4357da",
78 "metadata": {},
79 "outputs": [
80 {
81 "name": "stdout",
82 "output_type": "stream",
83 "text": [
84 "((array([[[0, 0, 0, ..., 0, 0, 0],\n",
85 " [0, 0, 0, ..., 0, 0, 0],\n",
86 " [0, 0, 0, ..., 0, 0, 0],\n",
87 " ...,\n",
88 " [0, 0, 0, ..., 0, 0, 0],\n",
89 " [0, 0, 0, ..., 0, 0, 0],\n",
90 " [0, 0, 0, ..., 0, 0, 0]],\n",
91 "\n",
92 " [[0, 0, 0, ..., 0, 0, 0],\n",
93 " [0, 0, 0, ..., 0, 0, 0],\n",
94 " [0, 0, 0, ..., 0, 0, 0],\n",
95 " ...,\n",
96 " [0, 0, 0, ..., 0, 0, 0],\n",
97 " [0, 0, 0, ..., 0, 0, 0],\n",
98 " [0, 0, 0, ..., 0, 0, 0]],\n",
99 "\n",
100 " [[0, 0, 0, ..., 0, 0, 0],\n",
101 " [0, 0, 0, ..., 0, 0, 0],\n",
102 " [0, 0, 0, ..., 0, 0, 0],\n",
103 " ...,\n",
104 " [0, 0, 0, ..., 0, 0, 0],\n",
105 " [0, 0, 0, ..., 0, 0, 0],\n",
106 " [0, 0, 0, ..., 0, 0, 0]],\n",
107 "\n",
108 " ...,\n",
109 "\n",
110 " [[0, 0, 0, ..., 0, 0, 0],\n",
111 " [0, 0, 0, ..., 0, 0, 0],\n",
112 " [0, 0, 0, ..., 0, 0, 0],\n",
113 " ...,\n",
114 " [0, 0, 0, ..., 0, 0, 0],\n",
115 " [0, 0, 0, ..., 0, 0, 0],\n",
116 " [0, 0, 0, ..., 0, 0, 0]],\n",
117 "\n",
118 " [[0, 0, 0, ..., 0, 0, 0],\n",
119 " [0, 0, 0, ..., 0, 0, 0],\n",
120 " [0, 0, 0, ..., 0, 0, 0],\n",
121 " ...,\n",
122 " [0, 0, 0, ..., 0, 0, 0],\n",
123 " [0, 0, 0, ..., 0, 0, 0],\n",
124 " [0, 0, 0, ..., 0, 0, 0]],\n",
125 "\n",
126 " [[0, 0, 0, ..., 0, 0, 0],\n",
127 " [0, 0, 0, ..., 0, 0, 0],\n",
128 " [0, 0, 0, ..., 0, 0, 0],\n",
129 " ...,\n",
130 " [0, 0, 0, ..., 0, 0, 0],\n",
131 " [0, 0, 0, ..., 0, 0, 0],\n",
132 " [0, 0, 0, ..., 0, 0, 0]]], dtype=uint8), array([5, 0, 4, ..., 5, 6, 8], dtype=uint8)), (array([[[0, 0, 0, ..., 0, 0, 0],\n",
133 " [0, 0, 0, ..., 0, 0, 0],\n",
134 " [0, 0, 0, ..., 0, 0, 0],\n",
135 " ...,\n",
136 " [0, 0, 0, ..., 0, 0, 0],\n",
137 " [0, 0, 0, ..., 0, 0, 0],\n",
138 " [0, 0, 0, ..., 0, 0, 0]],\n",
139 "\n",
140 " [[0, 0, 0, ..., 0, 0, 0],\n",
141 " [0, 0, 0, ..., 0, 0, 0],\n",
142 " [0, 0, 0, ..., 0, 0, 0],\n",
143 " ...,\n",
144 " [0, 0, 0, ..., 0, 0, 0],\n",
145 " [0, 0, 0, ..., 0, 0, 0],\n",
146 " [0, 0, 0, ..., 0, 0, 0]],\n",
147 "\n",
148 " [[0, 0, 0, ..., 0, 0, 0],\n",
149 " [0, 0, 0, ..., 0, 0, 0],\n",
150 " [0, 0, 0, ..., 0, 0, 0],\n",
151 " ...,\n",
152 " [0, 0, 0, ..., 0, 0, 0],\n",
153 " [0, 0, 0, ..., 0, 0, 0],\n",
154 " [0, 0, 0, ..., 0, 0, 0]],\n",
155 "\n",
156 " ...,\n",
157 "\n",
158 " [[0, 0, 0, ..., 0, 0, 0],\n",
159 " [0, 0, 0, ..., 0, 0, 0],\n",
160 " [0, 0, 0, ..., 0, 0, 0],\n",
161 " ...,\n",
162 " [0, 0, 0, ..., 0, 0, 0],\n",
163 " [0, 0, 0, ..., 0, 0, 0],\n",
164 " [0, 0, 0, ..., 0, 0, 0]],\n",
165 "\n",
166 " [[0, 0, 0, ..., 0, 0, 0],\n",
167 " [0, 0, 0, ..., 0, 0, 0],\n",
168 " [0, 0, 0, ..., 0, 0, 0],\n",
169 " ...,\n",
170 " [0, 0, 0, ..., 0, 0, 0],\n",
171 " [0, 0, 0, ..., 0, 0, 0],\n",
172 " [0, 0, 0, ..., 0, 0, 0]],\n",
173 "\n",
174 " [[0, 0, 0, ..., 0, 0, 0],\n",
175 " [0, 0, 0, ..., 0, 0, 0],\n",
176 " [0, 0, 0, ..., 0, 0, 0],\n",
177 " ...,\n",
178 " [0, 0, 0, ..., 0, 0, 0],\n",
179 " [0, 0, 0, ..., 0, 0, 0],\n",
180 " [0, 0, 0, ..., 0, 0, 0]]], dtype=uint8), array([7, 2, 1, ..., 4, 5, 6], dtype=uint8)))\n"
181 ]
182 }
183 ],
184 "source": [
185 "# You can preview the raw data prior to training the model\n",
186 "print(mnist.load_data())"
187 ]
188 },
189 {
190 "cell_type": "markdown",
191 "id": "82f07fd0-3341-4ac7-b2cc-ddc99802896f",
192 "metadata": {},
193 "source": [
194 "## Build a machine learning model"
195 ]
196 },
197 {
198 "cell_type": "code",
199 "execution_count": 7,
200 "id": "e3903d22-f584-4305-85a7-d7e1494cf909",
201 "metadata": {},
202 "outputs": [],
203 "source": [
204 "# Build a tf.keras.Sequential model:\n",
205 "model = tf.keras.models.Sequential([\n",
206 " tf.keras.layers.Flatten(input_shape=(28, 28)),\n",
207 " tf.keras.layers.Dense(128, activation='relu'),\n",
208 " tf.keras.layers.Dropout(0.2),\n",
209 " tf.keras.layers.Dense(10)\n",
210 "])"
211 ]
212 },
213 {
214 "cell_type": "code",
215 "execution_count": 8,
216 "id": "f4dd7df9-feb6-48b3-b332-4652812571d4",
217 "metadata": {},
218 "outputs": [
219 {
220 "data": {
221 "text/plain": [
222 "array([[ 0.28218323, -0.2626474 , -0.16938315, 0.15272117, -0.2957897 ,\n",
223 " -0.0528494 , 0.02909562, 0.06403146, 0.67431676, -0.35960984]],\n",
224 " dtype=float32)"
225 ]
226 },
227 "execution_count": 8,
228 "metadata": {},
229 "output_type": "execute_result"
230 }
231 ],
232 "source": [
233 "# For each example, the model returns a vector of logits or log-odds scores, one for each class.\n",
234 "predictions = model(x_train[:1]).numpy()\n",
235 "predictions"
236 ]
237 },
238 {
239 "cell_type": "code",
240 "execution_count": 9,
241 "id": "b9a5a663-8d95-4fc5-a569-efb6362454e9",
242 "metadata": {},
243 "outputs": [
244 {
245 "data": {
246 "text/plain": [
247 "array([[0.12565382, 0.07287167, 0.07999501, 0.1103954 , 0.07049612,\n",
248 " 0.08988202, 0.0975576 , 0.1010261 , 0.18598464, 0.0661376 ]],\n",
249 " dtype=float32)"
250 ]
251 },
252 "execution_count": 9,
253 "metadata": {},
254 "output_type": "execute_result"
255 }
256 ],
257 "source": [
258 "# The tf.nn.softmax function converts these logits to probabilities for each class: \n",
259 "tf.nn.softmax(predictions).numpy()"
260 ]
261 },
262 {
263 "cell_type": "code",
264 "execution_count": 10,
265 "id": "c11f1e4e-c6a8-4a65-a4cb-36486209797c",
266 "metadata": {},
267 "outputs": [],
268 "source": [
269 "# Define a loss function for training using losses.SparseCategoricalCrossentropy:\n",
270 "loss_fn = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True)"
271 ]
272 },
273 {
274 "cell_type": "code",
275 "execution_count": 12,
276 "id": "b23213f0-7818-4c58-9672-495bc6bd240a",
277 "metadata": {},
278 "outputs": [],
279 "source": [
280 "# Configure and compile the model\n",
281 "model.compile(optimizer='adam',\n",
282 " loss=loss_fn,\n",
283 " metrics=['accuracy'])\n"
284 ]
285 },
286 {
287 "cell_type": "markdown",
288 "id": "74edfcf4-7b45-407f-8523-40cfab8cabc7",
289 "metadata": {},
290 "source": [
291 "## Train and evaluate your model"
292 ]
293 },
294 {
295 "cell_type": "code",
296 "execution_count": 13,
297 "id": "c0438b2f-78f5-469f-a2f9-d31198ac6411",
298 "metadata": {},
299 "outputs": [
300 {
301 "name": "stdout",
302 "output_type": "stream",
303 "text": [
304 "Epoch 1/5\n",
305 "1875/1875 [==============================] - 1s 509us/step - loss: 0.3016 - accuracy: 0.9124\n",
306 "Epoch 2/5\n",
307 "1875/1875 [==============================] - 1s 514us/step - loss: 0.1462 - accuracy: 0.9572\n",
308 "Epoch 3/5\n",
309 "1875/1875 [==============================] - 1s 505us/step - loss: 0.1087 - accuracy: 0.9663\n",
310 "Epoch 4/5\n",
311 "1875/1875 [==============================] - 1s 512us/step - loss: 0.0893 - accuracy: 0.9718\n",
312 "Epoch 5/5\n",
313 "1875/1875 [==============================] - 1s 499us/step - loss: 0.0774 - accuracy: 0.9758\n"
314 ]
315 },
316 {
317 "data": {
318 "text/plain": [
319 "<keras.src.callbacks.History at 0x2977640d0>"
320 ]
321 },
322 "execution_count": 13,
323 "metadata": {},
324 "output_type": "execute_result"
325 }
326 ],
327 "source": [
328 "# Use the Model.fit method to adjust your model parameters and minimize the loss: \n",
329 "model.fit(x_train, y_train, epochs=5)"
330 ]
331 },
332 {
333 "cell_type": "code",
334 "execution_count": 14,
335 "id": "7778bac2-ebd4-43eb-94a8-03b79152c58a",
336 "metadata": {},
337 "outputs": [
338 {
339 "name": "stdout",
340 "output_type": "stream",
341 "text": [
342 "313/313 - 0s - loss: 0.0790 - accuracy: 0.9757 - 126ms/epoch - 403us/step\n"
343 ]
344 },
345 {
346 "data": {
347 "text/plain": [
348 "[0.07904709875583649, 0.9757000207901001]"
349 ]
350 },
351 "execution_count": 14,
352 "metadata": {},
353 "output_type": "execute_result"
354 }
355 ],
356 "source": [
357 "# The Model.evaluate method checks the model's performance, usually on a validation set or test set.\n",
358 "model.evaluate(x_test, y_test, verbose=2)"
359 ]
360 },
361 {
362 "cell_type": "code",
363 "execution_count": 15,
364 "id": "96d86df4-4f22-4d76-ac4b-720192239015",
365 "metadata": {},
366 "outputs": [
367 {
368 "data": {
369 "text/plain": [
370 "<tf.Tensor: shape=(5, 10), dtype=float32, numpy=\n",
371 "array([[2.1381442e-07, 1.2493059e-08, 4.6679975e-06, 5.0975598e-04,\n",
372 " 2.3767580e-10, 8.8744054e-07, 5.8283575e-13, 9.9947828e-01,\n",
373 " 4.8932998e-07, 5.7814891e-06],\n",
374 " [1.6270951e-08, 3.1651885e-05, 9.9994957e-01, 1.1931742e-05,\n",
375 " 4.2942398e-15, 9.1026629e-07, 1.0364544e-06, 8.1141607e-17,\n",
376 " 4.9234400e-06, 7.9949551e-15],\n",
377 " [1.7301611e-06, 9.9930012e-01, 5.5941098e-05, 2.8840779e-05,\n",
378 " 8.1860111e-05, 3.5271249e-05, 8.1873928e-05, 2.4437119e-04,\n",
379 " 1.6496866e-04, 5.0269696e-06],\n",
380 " [9.9992669e-01, 4.8858471e-08, 1.1441392e-05, 2.0616257e-07,\n",
381 " 5.4289058e-07, 6.2358333e-07, 7.2935950e-06, 5.1983669e-05,\n",
382 " 3.5523688e-09, 1.0397144e-06],\n",
383 " [4.4057975e-07, 7.7009216e-10, 7.9363446e-07, 5.2758939e-08,\n",
384 " 9.9748683e-01, 9.6599024e-08, 8.6932334e-07, 1.1146701e-05,\n",
385 " 5.3453311e-07, 2.4991999e-03]], dtype=float32)>"
386 ]
387 },
388 "execution_count": 15,
389 "metadata": {},
390 "output_type": "execute_result"
391 }
392 ],
393 "source": [
394 "# If you want your model to return a probability, you can wrap the trained model, and attach the softmax to it:\n",
395 "\n",
396 "probability_model = tf.keras.Sequential([\n",
397 " model,\n",
398 " tf.keras.layers.Softmax()\n",
399 "])\n",
400 "probability_model(x_test[:5])"
401 ]
402 },
403 {
404 "cell_type": "code",
405 "execution_count": null,
406 "id": "ab098ffa-ab7d-4a76-90e0-255aa0763d22",
407 "metadata": {},
408 "outputs": [],
409 "source": []
410 }
411 ],
412 "metadata": {
413 "kernelspec": {
414 "display_name": "Python 3 (ipykernel)",
415 "language": "python",
416 "name": "python3"
417 },
418 "language_info": {
419 "codemirror_mode": {
420 "name": "ipython",
421 "version": 3
422 },
423 "file_extension": ".py",
424 "mimetype": "text/x-python",
425 "name": "python",
426 "nbconvert_exporter": "python",
427 "pygments_lexer": "ipython3",
428 "version": "3.11.5"
429 }
430 },
431 "nbformat": 4,
432 "nbformat_minor": 5
433}