From 25b5d6983d51e0e791b96a76beb7e49f392cd9a8 Mon Sep 17 00:00:00 2001 From: aktersnurra Date: Mon, 7 Dec 2020 22:54:04 +0100 Subject: Segmentation working! --- src/notebooks/05a-UNet.ipynb | 413 +++++++++++++++++++++++++++++-------------- 1 file changed, 280 insertions(+), 133 deletions(-) (limited to 'src/notebooks/05a-UNet.ipynb') diff --git a/src/notebooks/05a-UNet.ipynb b/src/notebooks/05a-UNet.ipynb index c25865a..77d895d 100644 --- a/src/notebooks/05a-UNet.ipynb +++ b/src/notebooks/05a-UNet.ipynb @@ -23,76 +23,7 @@ }, { "cell_type": "code", - "execution_count": 5, - "metadata": {}, - "outputs": [], - "source": [ - "x = 64\n", - "depth = 4\n", - "channels = [x * 2 ** i for i in range(4)]" - ] - }, - { - "cell_type": "code", - "execution_count": 6, - "metadata": {}, - "outputs": [], - "source": [ - "channels.reverse()" - ] - }, - { - "cell_type": "code", - "execution_count": 9, - "metadata": {}, - "outputs": [ - { - "data": { - "text/plain": [ - "[512, 256, 128, 64]" - ] - }, - "execution_count": 9, - "metadata": {}, - "output_type": "execute_result" - } - ], - "source": [ - "channels" - ] - }, - { - "cell_type": "code", - "execution_count": 11, - "metadata": {}, - "outputs": [], - "source": [ - "m = nn.ModuleList([nn.Conv2d(1,3,2), nn.Linear(1, 5)])" - ] - }, - { - "cell_type": "code", - "execution_count": 12, - "metadata": {}, - "outputs": [ - { - "ename": "ModuleAttributeError", - "evalue": "'ModuleList' object has no attribute 'reverse'", - "output_type": "error", - "traceback": [ - "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m", - "\u001b[0;31mModuleAttributeError\u001b[0m Traceback (most recent call last)", - "\u001b[0;32m\u001b[0m in \u001b[0;36m\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mm\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mreverse\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m", - "\u001b[0;32m~/Library/Caches/pypoetry/virtualenvs/text-recognizer-cxOiES-R-py3.8/lib/python3.8/site-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36m__getattr__\u001b[0;34m(self, name)\u001b[0m\n\u001b[1;32m 769\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mname\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mmodules\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 770\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mmodules\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mname\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 771\u001b[0;31m raise ModuleAttributeError(\"'{}' object has no attribute '{}'\".format(\n\u001b[0m\u001b[1;32m 772\u001b[0m type(self).__name__, name))\n\u001b[1;32m 773\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n", - "\u001b[0;31mModuleAttributeError\u001b[0m: 'ModuleList' object has no attribute 'reverse'" - ] - } - ], - "source": [] - }, - { - "cell_type": "code", - "execution_count": 40, + "execution_count": 2, "metadata": {}, "outputs": [], "source": [ @@ -101,7 +32,7 @@ }, { "cell_type": "code", - "execution_count": 99, + "execution_count": 3, "metadata": {}, "outputs": [], "source": [ @@ -110,7 +41,7 @@ }, { "cell_type": "code", - "execution_count": 100, + "execution_count": 4, "metadata": {}, "outputs": [], "source": [ @@ -119,72 +50,68 @@ }, { "cell_type": "code", - "execution_count": 101, + "execution_count": 5, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "ModuleList(\n", - " (0): DownSamplingBlock(\n", - " (conv_block): ConvBlock(\n", - " (activation): ReLU(inplace=True)\n", - " (block): Sequential(\n", - " (0): Conv2d(1, 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", + " (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): DownSamplingBlock(\n", - " (conv_block): ConvBlock(\n", - " (activation): ReLU(inplace=True)\n", - " (block): Sequential(\n", - " (0): Conv2d(64, 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", + " (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): DownSamplingBlock(\n", - " (conv_block): ConvBlock(\n", - " (activation): ReLU(inplace=True)\n", - " (block): Sequential(\n", - " (0): Conv2d(128, 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", + " (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): DownSamplingBlock(\n", - " (conv_block): ConvBlock(\n", - " (activation): ReLU(inplace=True)\n", - " (block): Sequential(\n", - " (0): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n", - " (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n", - " (2): ReLU(inplace=True)\n", - " (3): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n", - " (4): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n", - " (5): ReLU(inplace=True)\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": 101, + "execution_count": 5, "metadata": {}, "output_type": "execute_result" } @@ -195,15 +122,15 @@ }, { "cell_type": "code", - "execution_count": 102, + "execution_count": 6, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "ModuleList(\n", - " (0): UpSamplingBlock(\n", - " (conv_block): ConvBlock(\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", @@ -216,8 +143,8 @@ " )\n", " (up_sampling): Upsample(scale_factor=2.0, mode=bilinear)\n", " )\n", - " (1): UpSamplingBlock(\n", - " (conv_block): ConvBlock(\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", @@ -230,8 +157,8 @@ " )\n", " (up_sampling): Upsample(scale_factor=2.0, mode=bilinear)\n", " )\n", - " (2): UpSamplingBlock(\n", - " (conv_block): ConvBlock(\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", @@ -247,7 +174,7 @@ ")" ] }, - "execution_count": 102, + "execution_count": 6, "metadata": {}, "output_type": "execute_result" } @@ -258,7 +185,7 @@ }, { "cell_type": "code", - "execution_count": 104, + "execution_count": 7, "metadata": {}, "outputs": [ { @@ -267,7 +194,7 @@ "Conv2d(64, 3, kernel_size=(1, 1), stride=(1, 1))" ] }, - "execution_count": 104, + "execution_count": 7, "metadata": {}, "output_type": "execute_result" } @@ -278,7 +205,25 @@ }, { "cell_type": "code", - "execution_count": 103, + "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": [ { @@ -287,21 +232,223 @@ "torch.Size([1, 3, 256, 256])" ] }, - "execution_count": 103, + "execution_count": 9, "metadata": {}, "output_type": "execute_result" } ], "source": [ - "net(x).shape" + "yy.shape" ] }, { "cell_type": "code", - "execution_count": null, + "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": [] + "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", -- cgit v1.2.3-70-g09d2