{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "b2d6f096-6123-4dd0-ae4f-c9abf70889c4",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-04-04T08:47:31.680497Z",
     "iopub.status.busy": "2024-04-04T08:47:31.678982Z",
     "iopub.status.idle": "2024-04-04T08:47:52.012034Z",
     "shell.execute_reply": "2024-04-04T08:47:52.010508Z",
     "shell.execute_reply.started": "2024-04-04T08:47:31.680435Z"
    }
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f0c8978e9991468cb37176b7ee5e7f40",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Downloading builder script:   0%|          | 0.00/1.63k [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "1821c061525d4be1b088a0b1624ba0f7",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Downloading metadata:   0%|          | 0.00/1.01k [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading and preparing dataset mnist/mnist (download: 11.06 MiB, generated: 19.44 MiB, post-processed: Unknown size, total: 30.50 MiB) to /root/.cache/huggingface/datasets/mnist/mnist/1.0.0/fda16c03c4ecfb13f165ba7e29cf38129ce035011519968cdaf74894ce91c9d4...\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "eb3da6f2299d4012862a28dd3a74565d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Downloading data files:   0%|          | 0/4 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "e53657e2c6ad4c108ad703dae1f87efe",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Downloading data:   0%|          | 0.00/9.91M [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2407ea55dc5348619ad666a16a3b97ff",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Downloading data:   0%|          | 0.00/28.9k [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "04a14bdbb03044979afd7c44ba0f424e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Downloading data:   0%|          | 0.00/1.65M [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "976dba4172974f01ae37ad90b1c6cb1d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Downloading data:   0%|          | 0.00/4.54k [00:00<?, ?B/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "56d60d708f2549a5a0c6acc0e2499884",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Extracting data files:   0%|          | 0/4 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "87ba67ec7d834407afd7fcf7a9d6ac13",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Generating train split:   0%|          | 0/60000 [00:00<?, ? examples/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "696c5019b4c24f049ce78744c1fecded",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Generating test split:   0%|          | 0/10000 [00:00<?, ? examples/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dataset mnist downloaded and prepared to /root/.cache/huggingface/datasets/mnist/mnist/1.0.0/fda16c03c4ecfb13f165ba7e29cf38129ce035011519968cdaf74894ce91c9d4. Subsequent calls will reuse this data.\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d5193a728c974100b02fcb82ac236d80",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/2 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from datasets import load_dataset\n",
    "\n",
    "dataset = load_dataset(\"mnist\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "7c5ea43d-6a70-4deb-a90e-1d7758c961a3",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-04-04T08:47:52.015528Z",
     "iopub.status.busy": "2024-04-04T08:47:52.014998Z",
     "iopub.status.idle": "2024-04-04T08:47:56.923333Z",
     "shell.execute_reply": "2024-04-04T08:47:56.922341Z",
     "shell.execute_reply.started": "2024-04-04T08:47:52.015483Z"
    }
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from PIL import Image\n",
    "import io\n",
    "import tensorflow as tf\n",
    "\n",
    "def convert_image(byte_like_str_image):\n",
    "    return tf.convert_to_tensor(np.asarray(Image.open(io.BytesIO(byte_like_str_image['bytes']))), dtype = tf.float32)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "daafde17-8100-4f49-b27e-8aad43b129c6",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-04-04T08:47:56.925474Z",
     "iopub.status.busy": "2024-04-04T08:47:56.924961Z",
     "iopub.status.idle": "2024-04-04T08:48:32.236745Z",
     "shell.execute_reply": "2024-04-04T08:48:32.234178Z",
     "shell.execute_reply.started": "2024-04-04T08:47:56.925446Z"
    }
   },
   "outputs": [],
   "source": [
    "from keras.utils import to_categorical\n",
    "\n",
    "dataset_train = dataset['train'].to_pandas()\n",
    "dataset_train['image'] = dataset_train['image'].map(convert_image)\n",
    "\n",
    "dataset_test = dataset['test'].to_pandas()\n",
    "dataset_test['image'] = dataset_test['image'].map(convert_image)\n",
    "\n",
    "# Convert labels to NumPy arrays\n",
    "X_train = np.array(dataset_train['image'].tolist())\n",
    "y_train = np.array(dataset_train['label'])\n",
    "\n",
    "X_test = np.array(dataset_test['image'].tolist())\n",
    "y_test = np.array(dataset_test['label'])\n",
    "# dataset_train['label'] = dataset_train['label'].astype('float32')\n",
    "# dataset_test['label'] = dataset_test['label'].astype('float32')\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "72022fd2-000d-4d5c-88d5-9afc62c283d5",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-04-04T08:48:32.239111Z",
     "iopub.status.busy": "2024-04-04T08:48:32.238769Z",
     "iopub.status.idle": "2024-04-04T08:48:32.369358Z",
     "shell.execute_reply": "2024-04-04T08:48:32.367989Z",
     "shell.execute_reply.started": "2024-04-04T08:48:32.239071Z"
    }
   },
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "\n",
    "model = tf.keras.models.Sequential([\n",
    "  tf.keras.layers.Flatten(input_shape=(28, 28)),\n",
    "  tf.keras.layers.Dense(128, activation='relu'),\n",
    "  tf.keras.layers.Dense(10)\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "dd7871ac-cacd-4866-bdda-67651f592262",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-04-04T08:48:32.371520Z",
     "iopub.status.busy": "2024-04-04T08:48:32.371132Z",
     "iopub.status.idle": "2024-04-04T08:48:32.391109Z",
     "shell.execute_reply": "2024-04-04T08:48:32.390162Z",
     "shell.execute_reply.started": "2024-04-04T08:48:32.371484Z"
    }
   },
   "outputs": [],
   "source": [
    "model.compile(\n",
    "    optimizer=tf.keras.optimizers.Adam(0.001),\n",
    "    loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),\n",
    "    metrics=[tf.keras.metrics.SparseCategoricalAccuracy()],\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "280e0d9d-d9e8-41d9-b9ad-666e84fc0bfa",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-04-04T08:48:32.392690Z",
     "iopub.status.busy": "2024-04-04T08:48:32.392405Z",
     "iopub.status.idle": "2024-04-04T08:49:02.402966Z",
     "shell.execute_reply": "2024-04-04T08:49:02.402204Z",
     "shell.execute_reply.started": "2024-04-04T08:48:32.392662Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "1875/1875 [==============================] - 7s 3ms/step - loss: 2.6159 - sparse_categorical_accuracy: 0.8597 - val_loss: 0.5951 - val_sparse_categorical_accuracy: 0.8762\n",
      "Epoch 2/5\n",
      "1875/1875 [==============================] - 6s 3ms/step - loss: 0.4041 - sparse_categorical_accuracy: 0.9071 - val_loss: 0.4567 - val_sparse_categorical_accuracy: 0.9024\n",
      "Epoch 3/5\n",
      "1875/1875 [==============================] - 6s 3ms/step - loss: 0.2977 - sparse_categorical_accuracy: 0.9247 - val_loss: 0.3300 - val_sparse_categorical_accuracy: 0.9237\n",
      "Epoch 4/5\n",
      "1875/1875 [==============================] - 5s 3ms/step - loss: 0.2646 - sparse_categorical_accuracy: 0.9344 - val_loss: 0.2909 - val_sparse_categorical_accuracy: 0.9368\n",
      "Epoch 5/5\n",
      "1875/1875 [==============================] - 5s 3ms/step - loss: 0.2418 - sparse_categorical_accuracy: 0.9391 - val_loss: 0.2886 - val_sparse_categorical_accuracy: 0.9330\n",
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 60000 entries, 0 to 59999\n",
      "Data columns (total 2 columns):\n",
      " #   Column  Non-Null Count  Dtype \n",
      "---  ------  --------------  ----- \n",
      " 0   image   60000 non-null  object\n",
      " 1   label   60000 non-null  int64 \n",
      "dtypes: int64(1), object(1)\n",
      "memory usage: 937.6+ KB\n"
     ]
    }
   ],
   "source": [
    "# model.fit(\n",
    "#     dataset_train['image'], dataset_test['label'],\n",
    "#     epochs=10,\n",
    "#     validation_data=(dataset_test['image'], dataset_test['label']),\n",
    "# )\n",
    "# dataset_train['label'].head(50)\n",
    "model.fit(X_train, y_train, epochs=5, validation_data=(X_test, y_test))\n",
    "\n",
    "dataset_train.info()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "c7317f9a-14f4-4908-9895-8bc085900e28",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2024-04-04T08:49:35.659063Z",
     "iopub.status.busy": "2024-04-04T08:49:35.658449Z",
     "iopub.status.idle": "2024-04-04T08:49:36.588926Z",
     "shell.execute_reply": "2024-04-04T08:49:36.588088Z",
     "shell.execute_reply.started": "2024-04-04T08:49:35.659019Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "313/313 [==============================] - 1s 3ms/step - loss: 0.2886 - sparse_categorical_accuracy: 0.9330\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "0.9330000281333923"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "loss, accuracy = model.evaluate(X_test, y_test)\n",
    "accuracy"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.16"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}