mirror of https://github.com/cupcakearmy/mnist.git
training
This commit is contained in:
cupcakearmy
parent
727f750aec
commit
7bd67d4079
|
|
@ -0,0 +1,11 @@
|
|||
# Training code in python with keras
|
||||
|
||||
## Files
|
||||
|
||||
I included both the python notebook and the normal python code.
|
||||
|
||||
The code itself is the same in both files, just take the format you prefer.
|
||||
|
||||
## Training
|
||||
|
||||
The training is super simple. Just the basic hello world of machine learning. You can easily run it online in google colab for free on a gpu.
|
||||
|
|
@ -0,0 +1,337 @@
|
|||
{
|
||||
"nbformat": 4,
|
||||
"nbformat_minor": 0,
|
||||
"metadata": {
|
||||
"colab": {
|
||||
"name": "simple.ipynb",
|
||||
"provenance": [],
|
||||
"collapsed_sections": []
|
||||
},
|
||||
"kernelspec": {
|
||||
"name": "python3",
|
||||
"display_name": "Python 3"
|
||||
},
|
||||
"accelerator": "GPU"
|
||||
},
|
||||
"cells": [
|
||||
{
|
||||
"cell_type": "code",
|
||||
"metadata": {
|
||||
"id": "BeibK12dAiot",
|
||||
"colab_type": "code",
|
||||
"colab": {}
|
||||
},
|
||||
"source": [
|
||||
"from tensorflow.keras.datasets import mnist\n",
|
||||
"from tensorflow.keras.models import Sequential\n",
|
||||
"from tensorflow.keras.layers import Dense, Dropout, Flatten\n",
|
||||
"from tensorflow.keras.layers import Conv2D, MaxPooling2D\n",
|
||||
"from tensorflow.keras.utils import to_categorical"
|
||||
],
|
||||
"execution_count": 0,
|
||||
"outputs": []
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"metadata": {
|
||||
"id": "VAUkfCwTA_Km",
|
||||
"colab_type": "code",
|
||||
"outputId": "8dabe82c-7999-49d6-d675-20979c006f10",
|
||||
"colab": {
|
||||
"base_uri": "https://localhost:8080/",
|
||||
"height": 51
|
||||
}
|
||||
},
|
||||
"source": [
|
||||
"(x_train, y_train), (x_test, y_test) = mnist.load_data()"
|
||||
],
|
||||
"execution_count": 0,
|
||||
"outputs": [
|
||||
{
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/mnist.npz\n",
|
||||
"11493376/11490434 [==============================] - 0s 0us/step\n"
|
||||
],
|
||||
"name": "stdout"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"metadata": {
|
||||
"id": "apDIDKb-BEjA",
|
||||
"colab_type": "code",
|
||||
"outputId": "6c25331a-2276-4d3d-9739-3acfb601c0dc",
|
||||
"colab": {
|
||||
"base_uri": "https://localhost:8080/",
|
||||
"height": 51
|
||||
}
|
||||
},
|
||||
"source": [
|
||||
"# Reshaping for channels_last (tensorflow) with one channel\n",
|
||||
"size = 28\n",
|
||||
"print(x_train.shape, x_test.shape)\n",
|
||||
"x_train = x_train.reshape(len(x_train), size, size, 1).astype('float32')\n",
|
||||
"x_test = x_test.reshape(len(x_test), size, size, 1).astype('float32')\n",
|
||||
"print(x_train.shape, x_test.shape)"
|
||||
],
|
||||
"execution_count": 0,
|
||||
"outputs": [
|
||||
{
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"(60000, 28, 28) (10000, 28, 28)\n",
|
||||
"(60000, 28, 28, 1) (10000, 28, 28, 1)\n"
|
||||
],
|
||||
"name": "stdout"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"metadata": {
|
||||
"id": "RwTWdxVbB6Wq",
|
||||
"colab_type": "code",
|
||||
"outputId": "1a75bb86-8e7b-425f-c064-2f12c8615d8c",
|
||||
"colab": {
|
||||
"base_uri": "https://localhost:8080/",
|
||||
"height": 34
|
||||
}
|
||||
},
|
||||
"source": [
|
||||
"# Normalize\n",
|
||||
"upper = max(x_train.max(), x_test.max())\n",
|
||||
"lower = min(x_train.min(), x_test.min())\n",
|
||||
"print(f'Max: {upper} Min: {lower}')\n",
|
||||
"x_train /= upper\n",
|
||||
"x_test /= upper"
|
||||
],
|
||||
"execution_count": 0,
|
||||
"outputs": [
|
||||
{
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"Max: 255.0 Min: 0.0\n"
|
||||
],
|
||||
"name": "stdout"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"metadata": {
|
||||
"id": "lYQVq00XBacf",
|
||||
"colab_type": "code",
|
||||
"colab": {}
|
||||
},
|
||||
"source": [
|
||||
"total_classes = 10\n",
|
||||
"y_train = to_categorical(y_train, total_classes)\n",
|
||||
"y_test = to_categorical(y_test, total_classes)"
|
||||
],
|
||||
"execution_count": 0,
|
||||
"outputs": []
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"metadata": {
|
||||
"id": "Ad2UK8LkBhKe",
|
||||
"colab_type": "code",
|
||||
"colab": {}
|
||||
},
|
||||
"source": [
|
||||
"# Make the model\n",
|
||||
"model = Sequential()\n",
|
||||
"model.add(Conv2D(64, (3, 3), activation='relu', input_shape=(size,size, 1), data_format='channels_last'))\n",
|
||||
"model.add(Conv2D(32, (3, 3), activation='relu'))\n",
|
||||
"model.add(MaxPooling2D(pool_size=(2, 2)))\n",
|
||||
"model.add(Dropout(0.25))\n",
|
||||
"model.add(Flatten())\n",
|
||||
"model.add(Dense(128, activation='relu'))\n",
|
||||
"model.add(Dropout(0.5))\n",
|
||||
"model.add(Dense(total_classes, activation='softmax'))\n",
|
||||
"\n",
|
||||
"model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])"
|
||||
],
|
||||
"execution_count": 0,
|
||||
"outputs": []
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"metadata": {
|
||||
"id": "v4Vb1_CfEVIl",
|
||||
"colab_type": "code",
|
||||
"outputId": "f00919d7-56b6-4577-9c97-283a48cbf804",
|
||||
"colab": {
|
||||
"base_uri": "https://localhost:8080/",
|
||||
"height": 459
|
||||
}
|
||||
},
|
||||
"source": [
|
||||
"# Train\n",
|
||||
"model.fit(x_train, y_train,\n",
|
||||
" batch_size=32,\n",
|
||||
" epochs=12,\n",
|
||||
" verbose=True)"
|
||||
],
|
||||
"execution_count": 0,
|
||||
"outputs": [
|
||||
{
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"Train on 60000 samples\n",
|
||||
"Epoch 1/12\n",
|
||||
"60000/60000 [==============================] - 15s 256us/sample - loss: 0.2029 - acc: 0.9382\n",
|
||||
"Epoch 2/12\n",
|
||||
"60000/60000 [==============================] - 8s 133us/sample - loss: 0.0827 - acc: 0.9750\n",
|
||||
"Epoch 3/12\n",
|
||||
"60000/60000 [==============================] - 8s 134us/sample - loss: 0.0650 - acc: 0.9800\n",
|
||||
"Epoch 4/12\n",
|
||||
"60000/60000 [==============================] - 7s 123us/sample - loss: 0.0541 - acc: 0.9839\n",
|
||||
"Epoch 5/12\n",
|
||||
"60000/60000 [==============================] - 7s 123us/sample - loss: 0.0476 - acc: 0.9850\n",
|
||||
"Epoch 6/12\n",
|
||||
"60000/60000 [==============================] - 8s 130us/sample - loss: 0.0391 - acc: 0.9883\n",
|
||||
"Epoch 7/12\n",
|
||||
"60000/60000 [==============================] - 8s 128us/sample - loss: 0.0354 - acc: 0.9889\n",
|
||||
"Epoch 8/12\n",
|
||||
"60000/60000 [==============================] - 8s 128us/sample - loss: 0.0330 - acc: 0.9895\n",
|
||||
"Epoch 9/12\n",
|
||||
"60000/60000 [==============================] - 7s 122us/sample - loss: 0.0283 - acc: 0.9909\n",
|
||||
"Epoch 10/12\n",
|
||||
"60000/60000 [==============================] - 7s 122us/sample - loss: 0.0275 - acc: 0.9909\n",
|
||||
"Epoch 11/12\n",
|
||||
"60000/60000 [==============================] - 7s 122us/sample - loss: 0.0265 - acc: 0.9918\n",
|
||||
"Epoch 12/12\n",
|
||||
"60000/60000 [==============================] - 7s 123us/sample - loss: 0.0239 - acc: 0.9922\n"
|
||||
],
|
||||
"name": "stdout"
|
||||
},
|
||||
{
|
||||
"output_type": "execute_result",
|
||||
"data": {
|
||||
"text/plain": [
|
||||
"<tensorflow.python.keras.callbacks.History at 0x7fa5b027d278>"
|
||||
]
|
||||
},
|
||||
"metadata": {
|
||||
"tags": []
|
||||
},
|
||||
"execution_count": 11
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"metadata": {
|
||||
"id": "BybcqHVqEaQD",
|
||||
"colab_type": "code",
|
||||
"outputId": "9b5a41a0-9420-4387-8a01-5462c40878bd",
|
||||
"colab": {
|
||||
"base_uri": "https://localhost:8080/",
|
||||
"height": 51
|
||||
}
|
||||
},
|
||||
"source": [
|
||||
"score = model.evaluate(x_test, y_test, verbose=0)\n",
|
||||
"print('Test loss:', score[0])\n",
|
||||
"print('Test accuracy:', score[1])"
|
||||
],
|
||||
"execution_count": 0,
|
||||
"outputs": [
|
||||
{
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"Test loss: 0.028924251638433304\n",
|
||||
"Test accuracy: 0.9922\n"
|
||||
],
|
||||
"name": "stdout"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"metadata": {
|
||||
"id": "NNnMBuoURyaX",
|
||||
"colab_type": "code",
|
||||
"colab": {}
|
||||
},
|
||||
"source": [
|
||||
"# Save for keras\n",
|
||||
"model.save(\"model.h5\")"
|
||||
],
|
||||
"execution_count": 0,
|
||||
"outputs": []
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"metadata": {
|
||||
"id": "Ia7WzlH9_UxY",
|
||||
"colab_type": "code",
|
||||
"outputId": "1d61e90a-f2a8-4ef8-8df2-301d61a89c8d",
|
||||
"colab": {
|
||||
"base_uri": "https://localhost:8080/",
|
||||
"height": 513
|
||||
}
|
||||
},
|
||||
"source": [
|
||||
"!pip install tensorflowjs\n",
|
||||
"import tensorflowjs as tfjs\n",
|
||||
"\n",
|
||||
"# Save for the web\n",
|
||||
"tfjs.converters.save_keras_model(model, './js')"
|
||||
],
|
||||
"execution_count": 0,
|
||||
"outputs": [
|
||||
{
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"Requirement already satisfied: tensorflowjs in /usr/local/lib/python3.6/dist-packages (1.4.0)\n",
|
||||
"Requirement already satisfied: six==1.11.0 in /usr/local/lib/python3.6/dist-packages (from tensorflowjs) (1.11.0)\n",
|
||||
"Requirement already satisfied: tensorflow==1.15.0 in /usr/local/lib/python3.6/dist-packages (from tensorflowjs) (1.15.0)\n",
|
||||
"Requirement already satisfied: h5py==2.8.0 in /usr/local/lib/python3.6/dist-packages (from tensorflowjs) (2.8.0)\n",
|
||||
"Requirement already satisfied: gast==0.2.2 in /usr/local/lib/python3.6/dist-packages (from tensorflowjs) (0.2.2)\n",
|
||||
"Requirement already satisfied: tensorflow-hub==0.5.0 in /usr/local/lib/python3.6/dist-packages (from tensorflowjs) (0.5.0)\n",
|
||||
"Requirement already satisfied: numpy==1.16.4 in /usr/local/lib/python3.6/dist-packages (from tensorflowjs) (1.16.4)\n",
|
||||
"Requirement already satisfied: PyInquirer==1.0.3 in /usr/local/lib/python3.6/dist-packages (from tensorflowjs) (1.0.3)\n",
|
||||
"Requirement already satisfied: absl-py>=0.7.0 in /usr/local/lib/python3.6/dist-packages (from tensorflow==1.15.0->tensorflowjs) (0.9.0)\n",
|
||||
"Requirement already satisfied: protobuf>=3.6.1 in /usr/local/lib/python3.6/dist-packages (from tensorflow==1.15.0->tensorflowjs) (3.10.0)\n",
|
||||
"Requirement already satisfied: tensorflow-estimator==1.15.1 in /usr/local/lib/python3.6/dist-packages (from tensorflow==1.15.0->tensorflowjs) (1.15.1)\n",
|
||||
"Requirement already satisfied: wheel>=0.26 in /usr/local/lib/python3.6/dist-packages (from tensorflow==1.15.0->tensorflowjs) (0.33.6)\n",
|
||||
"Requirement already satisfied: termcolor>=1.1.0 in /usr/local/lib/python3.6/dist-packages (from tensorflow==1.15.0->tensorflowjs) (1.1.0)\n",
|
||||
"Requirement already satisfied: grpcio>=1.8.6 in /usr/local/lib/python3.6/dist-packages (from tensorflow==1.15.0->tensorflowjs) (1.15.0)\n",
|
||||
"Requirement already satisfied: astor>=0.6.0 in /usr/local/lib/python3.6/dist-packages (from tensorflow==1.15.0->tensorflowjs) (0.8.1)\n",
|
||||
"Requirement already satisfied: keras-applications>=1.0.8 in /usr/local/lib/python3.6/dist-packages (from tensorflow==1.15.0->tensorflowjs) (1.0.8)\n",
|
||||
"Requirement already satisfied: tensorboard<1.16.0,>=1.15.0 in /usr/local/lib/python3.6/dist-packages (from tensorflow==1.15.0->tensorflowjs) (1.15.0)\n",
|
||||
"Requirement already satisfied: google-pasta>=0.1.6 in /usr/local/lib/python3.6/dist-packages (from tensorflow==1.15.0->tensorflowjs) (0.1.8)\n",
|
||||
"Requirement already satisfied: opt-einsum>=2.3.2 in /usr/local/lib/python3.6/dist-packages (from tensorflow==1.15.0->tensorflowjs) (3.1.0)\n",
|
||||
"Requirement already satisfied: keras-preprocessing>=1.0.5 in /usr/local/lib/python3.6/dist-packages (from tensorflow==1.15.0->tensorflowjs) (1.1.0)\n",
|
||||
"Requirement already satisfied: wrapt>=1.11.1 in /usr/local/lib/python3.6/dist-packages (from tensorflow==1.15.0->tensorflowjs) (1.11.2)\n",
|
||||
"Requirement already satisfied: Pygments>=2.2.0 in /usr/local/lib/python3.6/dist-packages (from PyInquirer==1.0.3->tensorflowjs) (2.5.2)\n",
|
||||
"Requirement already satisfied: prompt-toolkit==1.0.14 in /usr/local/lib/python3.6/dist-packages (from PyInquirer==1.0.3->tensorflowjs) (1.0.14)\n",
|
||||
"Requirement already satisfied: regex>=2016.11.21 in /usr/local/lib/python3.6/dist-packages (from PyInquirer==1.0.3->tensorflowjs) (2019.12.20)\n",
|
||||
"Requirement already satisfied: setuptools in /usr/local/lib/python3.6/dist-packages (from protobuf>=3.6.1->tensorflow==1.15.0->tensorflowjs) (42.0.2)\n",
|
||||
"Requirement already satisfied: werkzeug>=0.11.15 in /usr/local/lib/python3.6/dist-packages (from tensorboard<1.16.0,>=1.15.0->tensorflow==1.15.0->tensorflowjs) (0.16.0)\n",
|
||||
"Requirement already satisfied: markdown>=2.6.8 in /usr/local/lib/python3.6/dist-packages (from tensorboard<1.16.0,>=1.15.0->tensorflow==1.15.0->tensorflowjs) (3.1.1)\n",
|
||||
"Requirement already satisfied: wcwidth in /usr/local/lib/python3.6/dist-packages (from prompt-toolkit==1.0.14->PyInquirer==1.0.3->tensorflowjs) (0.1.8)\n"
|
||||
],
|
||||
"name": "stdout"
|
||||
}
|
||||
]
|
||||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"metadata": {
|
||||
"id": "NriyjqVh_ge9",
|
||||
"colab_type": "code",
|
||||
"colab": {}
|
||||
},
|
||||
"source": [
|
||||
""
|
||||
],
|
||||
"execution_count": 0,
|
||||
"outputs": []
|
||||
}
|
||||
]
|
||||
}
|
||||
|
|
@ -0,0 +1,67 @@
|
|||
# -*- coding: utf-8 -*-
|
||||
"""simple.ipynb
|
||||
|
||||
Automatically generated by Colaboratory.
|
||||
|
||||
Original file is located at
|
||||
https://colab.research.google.com/drive/1CGpActfOTQuiUkle2q40rg8Bf7ijfyAU
|
||||
"""
|
||||
|
||||
from tensorflow.keras.datasets import mnist
|
||||
from tensorflow.keras.models import Sequential
|
||||
from tensorflow.keras.layers import Dense, Dropout, Flatten
|
||||
from tensorflow.keras.layers import Conv2D, MaxPooling2D
|
||||
from tensorflow.keras.utils import to_categorical
|
||||
|
||||
(x_train, y_train), (x_test, y_test) = mnist.load_data()
|
||||
|
||||
# Reshaping for channels_last (tensorflow) with one channel
|
||||
size = 28
|
||||
print(x_train.shape, x_test.shape)
|
||||
x_train = x_train.reshape(len(x_train), size, size, 1).astype('float32')
|
||||
x_test = x_test.reshape(len(x_test), size, size, 1).astype('float32')
|
||||
print(x_train.shape, x_test.shape)
|
||||
|
||||
# Normalize
|
||||
upper = max(x_train.max(), x_test.max())
|
||||
lower = min(x_train.min(), x_test.min())
|
||||
print(f'Max: {upper} Min: {lower}')
|
||||
x_train /= upper
|
||||
x_test /= upper
|
||||
|
||||
total_classes = 10
|
||||
y_train = to_categorical(y_train, total_classes)
|
||||
y_test = to_categorical(y_test, total_classes)
|
||||
|
||||
# Make the model
|
||||
model = Sequential()
|
||||
model.add(Conv2D(64, (3, 3), activation='relu', input_shape=(size,size, 1), data_format='channels_last'))
|
||||
model.add(Conv2D(32, (3, 3), activation='relu'))
|
||||
model.add(MaxPooling2D(pool_size=(2, 2)))
|
||||
model.add(Dropout(0.25))
|
||||
model.add(Flatten())
|
||||
model.add(Dense(128, activation='relu'))
|
||||
model.add(Dropout(0.5))
|
||||
model.add(Dense(total_classes, activation='softmax'))
|
||||
|
||||
model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])
|
||||
|
||||
# Train
|
||||
model.fit(x_train, y_train,
|
||||
batch_size=32,
|
||||
epochs=12,
|
||||
verbose=True)
|
||||
|
||||
score = model.evaluate(x_test, y_test, verbose=0)
|
||||
print('Test loss:', score[0])
|
||||
print('Test accuracy:', score[1])
|
||||
|
||||
# Save for keras
|
||||
model.save("model.h5")
|
||||
|
||||
!pip install tensorflowjs
|
||||
import tensorflowjs as tfjs
|
||||
|
||||
# Save for the web
|
||||
tfjs.converters.save_keras_model(model, './js')
|
||||
|
||||
Loading…
Reference in New Issue