Create CNNetwork

cnn.py

@@ -0,0 +1,66 @@
+from torch import nn
+from torchsummary import summary
+
+
+class CNNetwork(nn.Module):
+
+    def __init__(self):
+        super().__init__()
+        # 4 conv blocks / flatten / linear / softmax
+        self.conv1 = nn.Sequential(
+            nn.Conv2d(
+                in_channels=1,
+                out_channels=16,
+                kernel_size=3,
+                stride=1,
+                padding=2
+            ),
+            nn.ReLU(),
+            nn.MaxPool2d(kernel_size=2)
+        )
+        self.conv2 = nn.Sequential(
+            nn.Conv2d(
+                in_channels=16,
+                out_channels=32,
+                kernel_size=3,
+                stride=1,
+                padding=2
+            ),
+            nn.ReLU(),
+            nn.MaxPool2d(kernel_size=2)
+        )
+        self.conv3 = nn.Sequential(
+            nn.Conv2d(
+                in_channels=32,
+                out_channels=64,
+                kernel_size=3,
+                stride=1,
+                padding=2
+            ),
+            nn.ReLU(),
+            nn.MaxPool2d(kernel_size=2)
+        )
+        self.conv4 = nn.Sequential(
+            nn.Conv2d(
+                in_channels=64,
+                out_channels=128,
+                kernel_size=3,
+                stride=1,
+                padding=2
+            ),
+            nn.ReLU(),
+            nn.MaxPool2d(kernel_size=2)
+        )
+        self.flatten = nn.Flatten()
+        self.linear = nn.Linear(128 * 5 * 4, 10)
+        self.softmax = nn.Softmax(dim=1)
+
+    def forward(self, input_data):
+        x = self.conv1(input_data)
+        x = self.conv2(x)
+        x = self.conv3(x)
+        x = self.conv4(x)
+        x = self.flatten(x)
+        logits = self.linear(x)
+        predictions = self.softmax(logits)
+        return predictions

notebooks/playground.ipynb

@@ -3,7 +3,7 @@
   {
    "cell_type": "code",
    "execution_count": 8,
-   "id": "7f11e761",
+   "id": "46dbbffd",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -14,7 +14,7 @@
   {
    "cell_type": "code",
    "execution_count": 10,
-   "id": "f3deb79d",
+   "id": "56056453",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -24,30 +24,32 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 76,
-   "id": "eb9888a5",
+   "execution_count": 86,
+   "id": "b5cbac9d",
    "metadata": {},
    "outputs": [],
    "source": [
     "import os\n",
     "\n",
-    "import torch"
+    "import torch\n",
+    "from torchsummary import summary"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 77,
-   "id": "75440e63",
+   "execution_count": 85,
+   "id": "1970ad63",
    "metadata": {},
    "outputs": [],
    "source": [
-    "from dataset import *"
+    "from dataset import *\n",
+    "from cnn import CNNetwork"
    ]
   },
   {
    "cell_type": "code",
    "execution_count": 78,
-   "id": "5b51f712",
+   "id": "c28b0c5e",
    "metadata": {},
    "outputs": [
     {
@@ -69,7 +71,7 @@
   {
    "cell_type": "code",
    "execution_count": 80,
-   "id": "253f87d6",
+   "id": "69025839",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -85,7 +87,7 @@
   {
    "cell_type": "code",
    "execution_count": 81,
-   "id": "3d5c127a",
+   "id": "8dfbb1b4",
    "metadata": {},
    "outputs": [
     {
@@ -106,7 +108,7 @@
   {
    "cell_type": "code",
    "execution_count": 82,
-   "id": "cbac184f",
+   "id": "1071e53d",
    "metadata": {},
    "outputs": [
     {
@@ -134,7 +136,7 @@
   {
    "cell_type": "code",
    "execution_count": 83,
-   "id": "2bd8c582",
+   "id": "7a6d8133",
    "metadata": {},
    "outputs": [
     {
@@ -152,10 +154,56 @@
     "dataset[0][0].shape"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": 87,
+   "id": "4b8f75a0",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "----------------------------------------------------------------\n",
+      "        Layer (type)               Output Shape         Param #\n",
+      "================================================================\n",
+      "            Conv2d-1           [-1, 16, 66, 46]             160\n",
+      "              ReLU-2           [-1, 16, 66, 46]               0\n",
+      "         MaxPool2d-3           [-1, 16, 33, 23]               0\n",
+      "            Conv2d-4           [-1, 32, 35, 25]           4,640\n",
+      "              ReLU-5           [-1, 32, 35, 25]               0\n",
+      "         MaxPool2d-6           [-1, 32, 17, 12]               0\n",
+      "            Conv2d-7           [-1, 64, 19, 14]          18,496\n",
+      "              ReLU-8           [-1, 64, 19, 14]               0\n",
+      "         MaxPool2d-9             [-1, 64, 9, 7]               0\n",
+      "           Conv2d-10           [-1, 128, 11, 9]          73,856\n",
+      "             ReLU-11           [-1, 128, 11, 9]               0\n",
+      "        MaxPool2d-12            [-1, 128, 5, 4]               0\n",
+      "          Flatten-13                 [-1, 2560]               0\n",
+      "           Linear-14                   [-1, 10]          25,610\n",
+      "          Softmax-15                   [-1, 10]               0\n",
+      "================================================================\n",
+      "Total params: 122,762\n",
+      "Trainable params: 122,762\n",
+      "Non-trainable params: 0\n",
+      "----------------------------------------------------------------\n",
+      "Input size (MB): 0.01\n",
+      "Forward/backward pass size (MB): 1.83\n",
+      "Params size (MB): 0.47\n",
+      "Estimated Total Size (MB): 2.31\n",
+      "----------------------------------------------------------------\n"
+     ]
+    }
+   ],
+   "source": [
+    "cnn = CNNetwork()\n",
+    "summary(cnn, (1, 64, 44))"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "c3c7b1d4",
+   "id": "888726ed",
    "metadata": {},
    "outputs": [],
    "source": []