diff --git a/Jupyter_Notebooks/Chapter_03_Neural_Networks/Improving_NNs.ipynb b/Jupyter_Notebooks/Chapter_03_Neural_Networks/Improving_NNs.ipynb index 7fe6e33f..9b870c53 100644 --- a/Jupyter_Notebooks/Chapter_03_Neural_Networks/Improving_NNs.ipynb +++ b/Jupyter_Notebooks/Chapter_03_Neural_Networks/Improving_NNs.ipynb @@ -125,7 +125,7 @@ "colab": { "base_uri": "https://localhost:8080/" }, - "outputId": "c4927ffc-463c-4874-b355-87c6e8764b37", + "outputId": "b1858005-9fea-4912-85fb-eec9e8b8337f", "trusted": true, "execution": { "iopub.status.busy": "2024-11-05T00:47:36.188858Z", @@ -144,7 +144,7 @@ ] } ], - "execution_count": 1 + "execution_count": 23 }, { "cell_type": "code", @@ -183,7 +183,7 @@ } }, "outputs": [], - "execution_count": 2 + "execution_count": 24 }, { "cell_type": "code", @@ -267,7 +267,7 @@ } }, "outputs": [], - "execution_count": 3 + "execution_count": 25 }, { "cell_type": "markdown", @@ -306,7 +306,7 @@ "base_uri": "https://localhost:8080/" }, "id": "BBPItL0-O08w", - "outputId": "29e35174-7412-4c45-ee6b-db1dd62457c8", + "outputId": "559c9fcc-b32b-422e-e75d-93c2c32f60a6", "trusted": true, "execution": { "iopub.status.busy": "2024-11-05T00:47:40.822542Z", @@ -328,7 +328,7 @@ ] } ], - "execution_count": 4 + "execution_count": 26 }, { "cell_type": "markdown", @@ -376,7 +376,53 @@ ] } ], - "execution_count": 5 + "execution_count": null + }, + { + "cell_type": "markdown", + "source": [ + "Let's Make Sense of the Data We Are Working on!" + ], + "metadata": { + "id": "s9mY0HHE9Xo0" + } + }, + { + "cell_type": "code", + "source": [ + "# We Will Show First 10 Images of the Train_dataset\n", + "\n", + "fig = plt.figure(figsize=(10, 10))\n", + "for i in range(10):\n", + " plt.subplot(5, 5, i+1)\n", + " plt.xticks([])\n", + " plt.yticks([])\n", + " plt.grid(False)\n", + " plt.imshow(train_dataset[i][0].squeeze(), cmap=plt.cm.gray)\n", + " plt.xlabel(f\"Label: {classes[train_dataset[i][1]]}\")\n", + "plt.show()" + ], + "metadata": { + "id": "W40_bX4N9WZG", + "outputId": "60b202c3-4990-4ec1-e64d-518dae71b3b2", + "colab": { + "base_uri": "https://localhost:8080/", + "height": 348 + } + }, + "execution_count": 27, + "outputs": [ + { + "output_type": "display_data", + "data": { + "text/plain": [ + "
" + ], + "image/png": "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\n" + }, + "metadata": {} + } + ] }, { "cell_type": "markdown", @@ -422,7 +468,7 @@ } }, "outputs": [], - "execution_count": 6 + "execution_count": null }, { "cell_type": "markdown", @@ -477,12 +523,14 @@ "metadata": {} } ], - "execution_count": 7 + "execution_count": null }, { "cell_type": "markdown", "source": [ - "It appears this dataset is far more challenging than the original `MNIST` digits dataset.\n" + "It appears this dataset is far more challenging than the original `MNIST` digits dataset.\n", + "\n", + "As is evident from the classes, they are more labels. Moreover, some images are rotated, i.e the letters 'a' & 'k'. So we must come up with a more robust and better model!" ], "metadata": { "id": "0r4dUDzsv9EG" @@ -551,7 +599,7 @@ "metadata": {} } ], - "execution_count": 8 + "execution_count": null }, { "cell_type": "markdown", @@ -615,7 +663,7 @@ } }, "outputs": [], - "execution_count": 9 + "execution_count": null }, { "cell_type": "code", @@ -661,7 +709,7 @@ "metadata": {} } ], - "execution_count": 10 + "execution_count": null }, { "cell_type": "markdown", @@ -729,7 +777,7 @@ } }, "outputs": [], - "execution_count": 11 + "execution_count": null }, { "cell_type": "code", @@ -775,7 +823,7 @@ "metadata": {} } ], - "execution_count": 12 + "execution_count": null }, { "cell_type": "markdown", @@ -827,7 +875,7 @@ "metadata": { "id": "RTaeIe_e8u38" }, - "execution_count": 13, + "execution_count": null, "outputs": [] }, { @@ -847,7 +895,7 @@ "height": 544 } }, - "execution_count": 14, + "execution_count": null, "outputs": [ { "output_type": "stream", @@ -914,7 +962,7 @@ "height": 463 } }, - "execution_count": 15, + "execution_count": null, "outputs": [ { "output_type": "stream", @@ -976,4 +1024,4 @@ } } ] -} +} \ No newline at end of file