From 7e8e54e84c63171e748bbf09516fd517e6821ace Mon Sep 17 00:00:00 2001 From: Gustaf Rydholm Date: Sat, 20 Mar 2021 18:09:06 +0100 Subject: Inital commit for refactoring to lightning --- src/notebooks/05a-UNet.ipynb | 482 ------------------------------------------- 1 file changed, 482 deletions(-) delete mode 100644 src/notebooks/05a-UNet.ipynb (limited to 'src/notebooks/05a-UNet.ipynb') diff --git a/src/notebooks/05a-UNet.ipynb b/src/notebooks/05a-UNet.ipynb deleted file mode 100644 index 77d895d..0000000 --- a/src/notebooks/05a-UNet.ipynb +++ /dev/null @@ -1,482 +0,0 @@ -{ - "cells": [ - { - "cell_type": "code", - "execution_count": 1, - "metadata": {}, - "outputs": [], - "source": [ - "%load_ext autoreload\n", - "%autoreload 2\n", - "\n", - "%matplotlib inline\n", - "import matplotlib.pyplot as plt\n", - "import numpy as np\n", - "from PIL import Image\n", - "import torch\n", - "from torch import nn\n", - "from importlib.util import find_spec\n", - "if find_spec(\"text_recognizer\") is None:\n", - " import sys\n", - " sys.path.append('..')" - ] - }, - { - "cell_type": "code", - "execution_count": 2, - "metadata": {}, - "outputs": [], - "source": [ - "from text_recognizer.networks.unet import UNet" - ] - }, - { - "cell_type": "code", - "execution_count": 3, - "metadata": {}, - "outputs": [], - "source": [ - "net = UNet()" - ] - }, - { - "cell_type": "code", - "execution_count": 4, - "metadata": {}, - "outputs": [], - "source": [ - "x = torch.rand(1, 1, 256, 256)" - ] - }, - { - "cell_type": "code", - "execution_count": 5, - "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "ModuleList(\n", - " (0): _DilationBlock(\n", - " (activation): ELU(alpha=1.0, inplace=True)\n", - " (conv): Sequential(\n", - " (0): Conv2d(1, 32, kernel_size=(5, 5), stride=(1, 1), padding=(6, 6), dilation=(3, 3))\n", - " (1): ELU(alpha=1.0, inplace=True)\n", - " )\n", - " (conv1): Sequential(\n", - " (0): Conv2d(1, 32, kernel_size=(1, 1), stride=(1, 1))\n", - " (1): ELU(alpha=1.0, inplace=True)\n", - " )\n", - " (bn): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n", - " (down_sampling): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n", - " )\n", - " (1): _DilationBlock(\n", - " (activation): ELU(alpha=1.0, inplace=True)\n", - " (conv): Sequential(\n", - " (0): Conv2d(64, 64, kernel_size=(5, 5), stride=(1, 1), padding=(6, 6), dilation=(3, 3))\n", - " (1): ELU(alpha=1.0, inplace=True)\n", - " )\n", - " (conv1): Sequential(\n", - " (0): Conv2d(64, 64, kernel_size=(1, 1), stride=(1, 1))\n", - " (1): ELU(alpha=1.0, inplace=True)\n", - " )\n", - " (bn): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n", - " (down_sampling): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n", - " )\n", - " (2): _DilationBlock(\n", - " (activation): ELU(alpha=1.0, inplace=True)\n", - " (conv): Sequential(\n", - " (0): Conv2d(128, 128, kernel_size=(5, 5), stride=(1, 1), padding=(6, 6), dilation=(3, 3))\n", - " (1): ELU(alpha=1.0, inplace=True)\n", - " )\n", - " (conv1): Sequential(\n", - " (0): Conv2d(128, 128, kernel_size=(1, 1), stride=(1, 1))\n", - " (1): ELU(alpha=1.0, inplace=True)\n", - " )\n", - " (bn): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n", - " (down_sampling): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n", - " )\n", - " (3): _DilationBlock(\n", - " (activation): ELU(alpha=1.0, inplace=True)\n", - " (conv): Sequential(\n", - " (0): Conv2d(256, 256, kernel_size=(5, 5), stride=(1, 1), padding=(6, 6), dilation=(3, 3))\n", - " (1): ELU(alpha=1.0, inplace=True)\n", - " )\n", - " (conv1): Sequential(\n", - " (0): Conv2d(256, 256, kernel_size=(1, 1), stride=(1, 1))\n", - " (1): ELU(alpha=1.0, inplace=True)\n", - " )\n", - " (bn): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n", - " )\n", - ")" - ] - }, - "execution_count": 5, - "metadata": {}, - "output_type": "execute_result" - } - ], - "source": [ - "net.encoder_blocks" - ] - }, - { - "cell_type": "code", - "execution_count": 6, - "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "ModuleList(\n", - " (0): _UpSamplingBlock(\n", - " (conv_block): _ConvBlock(\n", - " (activation): ReLU(inplace=True)\n", - " (block): Sequential(\n", - " (0): Conv2d(768, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n", - " (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n", - " (2): ReLU(inplace=True)\n", - " (3): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n", - " (4): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n", - " (5): ReLU(inplace=True)\n", - " )\n", - " )\n", - " (up_sampling): Upsample(scale_factor=2.0, mode=bilinear)\n", - " )\n", - " (1): _UpSamplingBlock(\n", - " (conv_block): _ConvBlock(\n", - " (activation): ReLU(inplace=True)\n", - " (block): Sequential(\n", - " (0): Conv2d(384, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n", - " (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n", - " (2): ReLU(inplace=True)\n", - " (3): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n", - " (4): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n", - " (5): ReLU(inplace=True)\n", - " )\n", - " )\n", - " (up_sampling): Upsample(scale_factor=2.0, mode=bilinear)\n", - " )\n", - " (2): _UpSamplingBlock(\n", - " (conv_block): _ConvBlock(\n", - " (activation): ReLU(inplace=True)\n", - " (block): Sequential(\n", - " (0): Conv2d(192, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n", - " (1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n", - " (2): ReLU(inplace=True)\n", - " (3): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n", - " (4): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n", - " (5): ReLU(inplace=True)\n", - " )\n", - " )\n", - " (up_sampling): Upsample(scale_factor=2.0, mode=bilinear)\n", - " )\n", - ")" - ] - }, - "execution_count": 6, - "metadata": {}, - "output_type": "execute_result" - } - ], - "source": [ - "net.decoder_blocks" - ] - }, - { - "cell_type": "code", - "execution_count": 7, - "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "Conv2d(64, 3, kernel_size=(1, 1), stride=(1, 1))" - ] - }, - "execution_count": 7, - "metadata": {}, - "output_type": "execute_result" - } - ], - "source": [ - "net.head" - ] - }, - { - "cell_type": "code", - "execution_count": 8, - "metadata": {}, - "outputs": [], - "source": [ - "yy = net(x)" - ] - }, - { - "cell_type": "code", - "execution_count": 19, - "metadata": {}, - "outputs": [], - "source": [ - "y = (torch.randn(1, 256, 256) > 0).long()" - ] - }, - { - "cell_type": "code", - "execution_count": 9, - "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "torch.Size([1, 3, 256, 256])" - ] - }, - "execution_count": 9, - "metadata": {}, - "output_type": "execute_result" - } - ], - "source": [ - "yy.shape" - ] - }, - { - "cell_type": "code", - "execution_count": 21, - "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "tensor([[[1, 0, 1, ..., 0, 1, 0],\n", - " [1, 0, 1, ..., 0, 1, 0],\n", - " [1, 1, 0, ..., 1, 1, 0],\n", - " ...,\n", - " [1, 0, 0, ..., 0, 1, 1],\n", - " [0, 0, 1, ..., 1, 1, 0],\n", - " [0, 0, 1, ..., 0, 0, 0]]])" - ] - }, - "execution_count": 21, - "metadata": {}, - "output_type": "execute_result" - } - ], - "source": [ - "y" - ] - }, - { - "cell_type": "code", - "execution_count": 54, - "metadata": { - "scrolled": true - }, - "outputs": [], - "source": [ - "loss = nn.CrossEntropyLoss()" - ] - }, - { - "cell_type": "code", - "execution_count": 55, - "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "tensor(1.2502, grad_fn=)" - ] - }, - "execution_count": 55, - "metadata": {}, - "output_type": "execute_result" - } - ], - "source": [ - "loss(yy, y)" - ] - }, - { - "cell_type": "code", - "execution_count": 10, - "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "tensor([[[[-0.1692, 0.1223, 0.1750, ..., -0.1869, -0.0585, 0.0462],\n", - " [-0.1302, -0.0230, 0.3185, ..., -0.3760, 0.0204, -0.0686],\n", - " [-0.1062, -0.0216, 0.4592, ..., 0.0990, 0.0808, -0.1419],\n", - " ...,\n", - " [ 0.1386, -0.2856, 0.3074, ..., -0.3874, -0.0322, 0.0503],\n", - " [ 0.3562, -0.0960, 0.0815, ..., 0.1893, 0.1438, 0.2804],\n", - " [-0.2106, -0.1988, 0.0016, ..., -0.0031, -0.2820, 0.0113]],\n", - "\n", - " [[-0.1542, -0.1322, -0.3917, ..., -0.2297, -0.2328, 0.0103],\n", - " [ 0.1040, 0.2189, -0.3661, ..., 0.4818, -0.3737, 0.1117],\n", - " [ 0.0735, -0.6487, -0.1899, ..., 0.2213, -0.1529, -0.1020],\n", - " ...,\n", - " [-0.2046, -0.1477, 0.2941, ..., 0.0652, -0.7276, 0.1676],\n", - " [ 0.0413, -0.2013, -0.3192, ..., -0.4947, -0.1179, -0.1000],\n", - " [-0.4108, 0.0199, 0.2238, ..., -0.4482, -0.2370, 0.0119]],\n", - "\n", - " [[ 0.0834, 0.1303, 0.0629, ..., 0.4766, -0.0481, 0.2538],\n", - " [ 0.1218, 0.1324, 0.2464, ..., 0.0081, 0.4444, 0.4583],\n", - " [ 0.1155, 0.1417, 0.2248, ..., 0.6365, -0.0040, 0.3144],\n", - " ...,\n", - " [ 0.0744, -0.0751, -0.5654, ..., -0.2890, -0.0437, 0.2719],\n", - " [ 0.1057, -0.1093, -0.3803, ..., 0.0229, 0.1403, 0.0944],\n", - " [-0.0958, -0.3931, -0.0186, ..., 0.2102, -0.0842, 0.1909]]]],\n", - " grad_fn=)" - ] - }, - "execution_count": 10, - "metadata": {}, - "output_type": "execute_result" - } - ], - "source": [ - "yy" - ] - }, - { - "cell_type": "code", - "execution_count": 39, - "metadata": {}, - "outputs": [], - "source": [ - "from torchsummary import summary" - ] - }, - { - "cell_type": "code", - "execution_count": 47, - "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "==========================================================================================\n", - "Layer (type:depth-idx) Output Shape Param #\n", - "==========================================================================================\n", - "├─ModuleList: 1 [] --\n", - "| └─DownSamplingBlock: 2-1 [-1, 64, 128, 128] --\n", - "| | └─ConvBlock: 3-1 [-1, 64, 256, 256] 37,824\n", - "| | └─MaxPool2d: 3-2 [-1, 64, 128, 128] --\n", - "| └─DownSamplingBlock: 2-2 [-1, 128, 64, 64] --\n", - "| | └─ConvBlock: 3-3 [-1, 128, 128, 128] 221,952\n", - "| | └─MaxPool2d: 3-4 [-1, 128, 64, 64] --\n", - "| └─DownSamplingBlock: 2-3 [-1, 256, 32, 32] --\n", - "| | └─ConvBlock: 3-5 [-1, 256, 64, 64] 886,272\n", - "| | └─MaxPool2d: 3-6 [-1, 256, 32, 32] --\n", - "| └─DownSamplingBlock: 2-4 [-1, 512, 32, 32] --\n", - "| | └─ConvBlock: 3-7 [-1, 512, 32, 32] 3,542,016\n", - "├─ModuleList: 1 [] --\n", - "| └─UpSamplingBlock: 2-5 [-1, 256, 64, 64] --\n", - "| | └─Upsample: 3-8 [-1, 512, 64, 64] --\n", - "| | └─ConvBlock: 3-9 [-1, 256, 64, 64] 2,360,832\n", - "| └─UpSamplingBlock: 2-6 [-1, 128, 128, 128] --\n", - "| | └─Upsample: 3-10 [-1, 256, 128, 128] --\n", - "| | └─ConvBlock: 3-11 [-1, 128, 128, 128] 590,592\n", - "| └─UpSamplingBlock: 2-7 [-1, 64, 256, 256] --\n", - "| | └─Upsample: 3-12 [-1, 128, 256, 256] --\n", - "| | └─ConvBlock: 3-13 [-1, 64, 256, 256] 147,840\n", - "├─Conv2d: 1-1 [-1, 3, 256, 256] 195\n", - "==========================================================================================\n", - "Total params: 7,787,523\n", - "Trainable params: 7,787,523\n", - "Non-trainable params: 0\n", - "Total mult-adds (M): 35.93\n", - "==========================================================================================\n", - "Input size (MB): 0.25\n", - "Forward/backward pass size (MB): 1.50\n", - "Params size (MB): 29.71\n", - "Estimated Total Size (MB): 31.46\n", - "==========================================================================================\n" - ] - }, - { - "data": { - "text/plain": [ - "==========================================================================================\n", - "Layer (type:depth-idx) Output Shape Param #\n", - "==========================================================================================\n", - "├─ModuleList: 1 [] --\n", - "| └─DownSamplingBlock: 2-1 [-1, 64, 128, 128] --\n", - "| | └─ConvBlock: 3-1 [-1, 64, 256, 256] 37,824\n", - "| | └─MaxPool2d: 3-2 [-1, 64, 128, 128] --\n", - "| └─DownSamplingBlock: 2-2 [-1, 128, 64, 64] --\n", - "| | └─ConvBlock: 3-3 [-1, 128, 128, 128] 221,952\n", - "| | └─MaxPool2d: 3-4 [-1, 128, 64, 64] --\n", - "| └─DownSamplingBlock: 2-3 [-1, 256, 32, 32] --\n", - "| | └─ConvBlock: 3-5 [-1, 256, 64, 64] 886,272\n", - "| | └─MaxPool2d: 3-6 [-1, 256, 32, 32] --\n", - "| └─DownSamplingBlock: 2-4 [-1, 512, 32, 32] --\n", - "| | └─ConvBlock: 3-7 [-1, 512, 32, 32] 3,542,016\n", - "├─ModuleList: 1 [] --\n", - "| └─UpSamplingBlock: 2-5 [-1, 256, 64, 64] --\n", - "| | └─Upsample: 3-8 [-1, 512, 64, 64] --\n", - "| | └─ConvBlock: 3-9 [-1, 256, 64, 64] 2,360,832\n", - "| └─UpSamplingBlock: 2-6 [-1, 128, 128, 128] --\n", - "| | └─Upsample: 3-10 [-1, 256, 128, 128] --\n", - "| | └─ConvBlock: 3-11 [-1, 128, 128, 128] 590,592\n", - "| └─UpSamplingBlock: 2-7 [-1, 64, 256, 256] --\n", - "| | └─Upsample: 3-12 [-1, 128, 256, 256] --\n", - "| | └─ConvBlock: 3-13 [-1, 64, 256, 256] 147,840\n", - "├─Conv2d: 1-1 [-1, 3, 256, 256] 195\n", - "==========================================================================================\n", - "Total params: 7,787,523\n", - "Trainable params: 7,787,523\n", - "Non-trainable params: 0\n", - "Total mult-adds (M): 35.93\n", - "==========================================================================================\n", - "Input size (MB): 0.25\n", - "Forward/backward pass size (MB): 1.50\n", - "Params size (MB): 29.71\n", - "Estimated Total Size (MB): 31.46\n", - "==========================================================================================" - ] - }, - "execution_count": 47, - "metadata": {}, - "output_type": "execute_result" - } - ], - "source": [ - "\n", - "summary(net, (1, 256, 256), device=\"cpu\")" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] - } - ], - "metadata": { - "kernelspec": { - "display_name": "Python 3", - "language": "python", - "name": "python3" - }, - "language_info": { - "codemirror_mode": { - "name": "ipython", - "version": 3 - }, - "file_extension": ".py", - "mimetype": "text/x-python", - "name": "python", - "nbconvert_exporter": "python", - "pygments_lexer": "ipython3", - "version": "3.8.2" - } - }, - "nbformat": 4, - "nbformat_minor": 4 -} -- cgit v1.2.3-70-g09d2