Etude_sur_Binary_alpha_digits.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd \n",
    "import matplotlib.pyplot as plt\n",
    "from tqdm import tqdm\n",
    "import scipy.io as sio\n",
    "\n",
    "from principal_RBM_alpha import init_RBM, train_RBM, lire_alpha_digit, generer_image_RBM\n",
    "from principal_DBN_alpha import init_DBN, train_DBN, generer_image_DBN\n",
    "import os"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Get the current working directory\n",
    "current_directory = os.getcwd()\n",
    "\n",
    "binary_alpha_digs_file = 'data\\\\binaryalphadigs.mat'\n",
    "\n",
    "binary_alpha_digs_path = os.path.join(current_directory, binary_alpha_digs_file)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 1. RBM analysis "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1.1 Number of neurons in hidden layers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "epochs = 100\n",
    "learning_rate = 0.01\n",
    "batch_size = 128\n",
    "characters_index = [2]\n",
    "neurons_list = [320, 16, 32, 64, 128, 256, 512]\n",
    "\n",
    "n_iter_gibs = 1000\n",
    "n_imgs = 10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training with epochs = 100 \n",
      "neurons = 16 \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training with epochs = 100 \n",
      "neurons = 32 \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training with epochs = 100 \n",
      "neurons = 64 \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training with epochs = 100 \n",
      "neurons = 128 \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training with epochs = 100 \n",
      "neurons = 256 \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training with epochs = 100 \n",
      "neurons = 512 \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "image = lire_alpha_digit(binary_alpha_digs_path, characters_index)\n",
    "\n",
    "all_generated_imgs = []\n",
    "\n",
    "for index in range(len(neurons_list) - 1) :\n",
    "    \n",
    "    W, a, b = init_RBM(neurons_list[0], neurons_list[index + 1])\n",
    "    W, a, b, loss = train_RBM(image, W, a, b, epochs, learning_rate, batch_size, verbose=False)\n",
    "\n",
    "    print('Training with epochs =' , epochs, '\\nneurons =', neurons_list[index + 1], '\\nlearning_rate = ', learning_rate, '\\nbatch_size = ' , batch_size, '\\non characters : ', characters_index)\n",
    "    \n",
    "    generated_imgs = generer_image_RBM(n_imgs, n_iter_gibs, W, a, b)\n",
    "    \n",
    "    all_generated_imgs.append(generated_imgs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1080x864 with 60 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "n_rows = len(neurons_list) - 1  \n",
    "n_cols = n_imgs\n",
    "\n",
    "# Further adjust the figure size if needed\n",
    "fig, axs = plt.subplots(n_rows, n_cols, figsize=(15, 2 * n_rows))\n",
    "\n",
    "# Ensure axs is 2D for consistent indexing\n",
    "axs = axs.reshape(n_rows, n_cols) if n_rows > 1 else np.expand_dims(axs, 0)\n",
    "\n",
    "fig.patch.set_facecolor('black')\n",
    "\n",
    "# Set spacing between subplots and adjust margins\n",
    "plt.subplots_adjust(left=0.2, right=0.9, top=0.85, bottom=0.1, hspace=0.3, wspace=0.2)\n",
    "\n",
    "for i in range(n_rows):\n",
    "    for j in range(n_cols):\n",
    "        axs[i, j].imshow(all_generated_imgs[i][j].reshape(20, 16), cmap='gray')\n",
    "        axs[i, j].axis('off')\n",
    "\n",
    "        # Adjust the neuron label positioning and font\n",
    "        if j == 0:  # Only for the first column\n",
    "            axs[i, j].text(-0.5, 0.5, f'Neurons: {neurons_list[i + 1]}', \n",
    "                           va='center', ha='left', rotation='vertical', \n",
    "                           fontsize=12, color=\"white\", fontweight='bold', transform=axs[i, j].transAxes)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1.2 Number of different characters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "epochs = 100\n",
    "learning_rate = 0.01\n",
    "batch_size = 128\n",
    "n_neurons = [320, 128]\n",
    "characters_list = [[2], [2, 3], [2, 3, 4], [2, 3, 4, 5], [2, 3, 4, 5, 7], [2, 3, 4, 5, 7, 9]]\n",
    "\n",
    "n_iter_gibs = 1000\n",
    "n_imgs = 10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "training with epochs = 100 \n",
      "neurons = [320, 128] \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "training with epochs = 100 \n",
      "neurons = [320, 128] \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2, 3]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "training with epochs = 100 \n",
      "neurons = [320, 128] \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2, 3, 4]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "training with epochs = 100 \n",
      "neurons = [320, 128] \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2, 3, 4, 5]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAp4AAAEYCAYAAAD1d/bLAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/H5lhTAAAACXBIWXMAAAsTAAALEwEAmpwYAAALHElEQVR4nO3d0W7jOBIFUHmR///l3qdFT2ZbaospXxbFc16DWLJYpi8IVPl1HMevAwAAPuw/s28AAIA9CJ4AAEQIngAARAieAABECJ4AAER8zb6BUb9+/bkZ//V6he+E7hK1cnaN6uvwOVdr+CfV63r3+lfU3OdU7ieVa35FPXw3Oz+MXH+kVrquuxNPAAAiBE8AACIETwAAIgRPAAAiBE8AACJeR+i32qu7flPdgHeNdKV17TzrKNUN2NXqtZLqzKy8zuzXGrHjPrT73jBi9TWvtns9HEemJpx4AgAQIXgCABAheAIAECF4AgAQIXgCABAheAIAENFinNLunjzSovP4nBU9uVbu2mXNq61cQ9a81sq18AnqyzglAAAeRPAEACBC8AQAIELwBAAgQvAEACBCV3vQWbfY1bPZsetwdq1cPfOzexv5n2or1Erl85s9/aD6eXe+ty5G9srqz9/d69hPnmH299JslfXgxBMAgAjBEwCACMETAIAIwRMAgAjBEwCACMETAICI1uOUUiNW7l7j6jrV75Ofm10rldTKb7uMIUuNA2Jt6uR9ie+EzmbvnU48AQCIEDwBAIgQPAEAiBA8AQCIEDwBAIiIdbWvSKcyFSrraPakh9XN7uasZooG7xjZG+wna1gxpzjxBAAgQvAEACBC8AQAIELwBAAgQvAEACBCV/sg3aS8S63wE+qHd/itdv6t697hxBMAgAjBEwCACMETAIAIwRMAgAjBEwCACMETAICIr9k30EH1GAr2Uz224uz1jDjhn9TDfiq/r9TP+lbML048AQCIEDwBAIgQPAEAiBA8AQCIEDwBAIh4XFf7SDfw2d+qO5XppXN3qDr6jKs1v7sPjKzRyPV3tPtUh+pO5V2eWzddO85n14MTTwAAIgRPAAAiBE8AACIETwAAIgRPAAAiBE8AACJex3H07Pf/C2MKeNfdWqkek1N5HfrosAepoc+oHLvVgTr5nM7rfmZ2PTjxBAAgQvAEACBC8AQAIELwBAAgQvAEACCidVd7Zadw586z2R1mT9d57e9SK59zViepKQeV1MnnzF7bSurks7rWytW6V+6DZ5x4AgAQIXgCABAheAIAECF4AgAQIXgCABCxTVd79XVSdB32MrvzeXY3Ir913jeuqIc+Vqwh9dNLhxq6WxNOPAEAiBA8AQCIEDwBAIgQPAEAiBA8AQCIEDwBAIhoPU6ps8oRBkbk7Kl6DIaaWNvssSjq5zOu1nXF8X/qZB2pWjFOCQCAlgRPAAAiBE8AACIETwAAIgRPAAAidLUX69pFxlpSUxNY20idqIf9jExHMVFlT4nvHieeAABECJ4AAEQIngAARAieAABECJ4AAEQIngAARHzNvoGnqR41kRrPRC9ndaQe+CljcvZjP+FdI/vA3Tpy4gkAQITgCQBAhOAJAECE4AkAQITgCQBAxFBX+1UHk87I+57cWThSK+oLqGI/OXf1/p/8vcRcTjwBAIgQPAEAiBA8AQCIEDwBAIgQPAEAiGjxW+1+O3hP1h1+xm9w/2Y/ue/JdTLy3tTKucpaceIJAECE4AkAQITgCQBAhOAJAECE4AkAQITgCQBARPk4JSMMzhld8fPXUivsRi38jO8kNVRhl3FdiVpx4gkAQITgCQBAhOAJAECE4AkAQITgCQBAxGVX++xOuNndiLPfP/9vpLPw7jqO1JBa+ZkVn19lzfFd6vmdrWHl9UfqJFVbT6jh1ESVytfb/TvGiScAABGCJwAAEYInAAARgicAABGCJwAAEYInAAARr+M4bvfoP6mtf1WVY6MqqIm+utXKn1SPTlOP51aohzOd6yQxmilplTpZ9fnu4KyGnHgCABAheAIAECF4AgAQIXgCABAheAIAEDHU1Q4AAHc58QQAIELwBAAgQvAEACBC8AQAIELwBAAgQvAEACBC8AQAIELwBAAgQvAEACBC8AQAIOJr9g108OtX3a+Gvl6v29e5+p8nq3wes9dw9PXoY3Y9Vl5HzX3O3bWtXouRNVcnP1f53M+kvnuuXitRE048AQCIEDwBAIgQPAEAiBA8AQCIEDwBAIh4Hcdxu1WqsoOYfbsRE12CnT1hDRN2WfPU+1y57p5UC0krr/kotVKrsoaceAIAECF4AgAQIXgCABAheAIAECF4AgAQIXgCABAxNE7pihEG9+046uKKGjqnVn5TJ2N2qyF1cu4JtbD7WL4U45QAAFiO4AkAQITgCQBAhOAJAECE4AkAQMTXyD+NdISNdERdXefs9SrvLXV9vhvpRqzsYNQN2cvsZ57Ya0Zfb+Q6T9W1Tq7MvucnSH1/z/5Mza6VymfjxBMAgAjBEwCACMETAIAIwRMAgAjBEwCACMETAICI13EckR79p402OJMaNcUcs0daXFFHz2RP+ZldvkeudH2f3axYK1e6jmhz4gkAQITgCQBAhOAJAECE4AkAQITgCQBARKyrfRe7dO9/0tOeR9fOQtZQ3dV+9npq67fZe1BqeoY1598S+4MTTwAAIgRPAAAiBE8AACIETwAAIgRPAAAidLUH6Sbdk672vBU/a+qEd+h4511d9xQnngAARAieAABECJ4AAEQIngAARAieAABECJ4AAER8zb6BUSuOS+l8b/zM1diKs3VPjUXZUdfPWnWddH2f/Ez13qBOPsfn834ec+IJAECE4AkAQITgCQBAhOAJAECE4AkAQMTrOI4tWmt1k/JJauW5VpygwdpGutrVI++a/X3lxBMAgAjBEwCACMETAIAIwRMAgAjBEwCACMETAICIZccp3R1xMjKeIsUYjM/quvbWvRd1wjsqx2tV1pw6WUfXveZKZX058QQAIELwBAAgQvAEACBC8AQAIELwBAAgonVX+4qdXyN0I/7c02pFTXzGVZ2sOBHjjPr5mRXX/Ip66GVkMsLsmtTVDgDAcgRPAAAiBE8AACIETwAAIgRPAAAiWne1zza7i+yKLsU51AT/U90hn+qqVyd/N7JOI6/VmTr5bqQTPXH9zs6ejRNPAAAiBE8AACIETwAAIgRPAAAiBE8AACIETwAAIoxTGjR7tIFRF+tI1Yqa6G/2vnFF/azNPrOnFfcUJ54AAEQIngAARAieAABECJ4AAEQIngAAROhqD6rsPtNZ+GxntXK17iP/w9rsKbxDnfCuRK048QQAIELwBAAgQvAEACBC8AQAIELwBAAgQvAEACDia/YN7ORstEDl+AKewcgSoMrIfrLK95IxcrUSOcWJJwAAEYInAAARgicAABGCJwAAEYInAAARutpp56p7bvdOxVU6TaGLkc/MLvvME/aTXdbqSZx4AgAQIXgCABAheAIAECF4AgAQIXgCABChq50Slb+Xq0uRf/JbzOcSv6vM+kY+Q2rouyftQ7MnxzjxBAAgQvAEACBC8AQAIELwBAAgQvAEACBC8AQAIGJonNLsVvzOjKD4LvU8Vqy76s/RCs+gsh5GXmuFZ3THbvuNNT838myeXD8j+2v1M5xde13X14knAAARgicAABGCJwAAEYInAAARgicAABGv4zhutz117ZTayexuuXd1qJXEs+rwPs90qpXOz+lM5fPr/P471cmZXZ7fLu+zQudndWbkGVZOdJg9bcaJJwAAEYInAAARgicAABGCJwAAEYInAAARgicAABHGKTVwNVrh7FkbafE5I+vRWbda+ZPKUSGjr7eLFerhjHXNWblO/mbFOpo9GmmEcUoAAEwleAIAECF4AgAQIXgCABAheAIAEDHU1Q4AAHc58QQAIELwBAAgQvAEACBC8AQAIELwBAAgQvAEACDivyrREQ7JkmRFAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "training with epochs = 100 \n",
      "neurons = [320, 128] \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2, 3, 4, 5, 7]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "training with epochs = 100 \n",
      "neurons = [320, 128] \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2, 3, 4, 5, 7, 9]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "all_generated_imgs = []\n",
    "\n",
    "for index in range(len(characters_list)):\n",
    "    \n",
    "    image = lire_alpha_digit(binary_alpha_digs_path, characters_list[index])\n",
    "    \n",
    "    W,a,b = init_RBM(n_neurons[0], n_neurons[1])\n",
    "    W,a,b,loss = train_RBM(image, W, a, b, epochs, learning_rate, batch_size, verbose= False)\n",
    "\n",
    "    print('Training with epochs =' , epochs, '\\nneurons =', n_neurons, '\\nlearning_rate = ', learning_rate, '\\nbatch_size = ' , batch_size, '\\non characters : ', characters_list[index])\n",
    "\n",
    "    generated_imgs = generer_image_RBM(n_imgs, n_iter_gibs, W, a, b)\n",
    "    \n",
    "    all_generated_imgs.append(generated_imgs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1080x864 with 60 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "n_rows = len(characters_list)  \n",
    "n_cols = n_imgs\n",
    "\n",
    "# Further adjust the figure size if needed\n",
    "fig, axs = plt.subplots(n_rows, n_cols, figsize=(15, 2 * n_rows))\n",
    "\n",
    "# Ensure axs is 2D for consistent indexing\n",
    "axs = axs.reshape(n_rows, n_cols) if n_rows > 1 else np.expand_dims(axs, 0)\n",
    "\n",
    "fig.patch.set_facecolor('black')\n",
    "\n",
    "# Set spacing between subplots and adjust margins\n",
    "plt.subplots_adjust(left=0.2, right=0.9, top=0.85, bottom=0.1, hspace=0.3, wspace=0.2)\n",
    "\n",
    "for i in range(n_rows):\n",
    "    for j in range(n_cols):\n",
    "        axs[i, j].imshow(all_generated_imgs[i][j].reshape(20, 16), cmap='gray')\n",
    "        axs[i, j].axis('off')\n",
    "\n",
    "        # Adjust the neuron label positioning and font\n",
    "        if j == 0:  # Only for the first column\n",
    "            axs[i, j].text(-0.5, 0.5, f'{characters_list[i]}', \n",
    "                           va='center', ha='left', rotation='vertical', \n",
    "                           fontsize=12, color=\"white\", fontweight='bold', transform=axs[i, j].transAxes)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 2. DBN analysis"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2.1 The number of neurons in hidden layers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "List of neurons: [[320, 16, 16, 16], [320, 32, 32, 32], [320, 64, 64, 64], [320, 128, 128, 128], [320, 256, 256, 256], [320, 512, 512, 512]]\n"
     ]
    }
   ],
   "source": [
    "input_neurons = 320\n",
    "min_neurons_per_layer = 16\n",
    "hidden_layer_count = 3\n",
    "max_power_of_two = 5\n",
    "\n",
    "neurons_list = []\n",
    "for exponent in range(max_power_of_two + 1):\n",
    "    hidden_layer_neurons = min_neurons_per_layer * 2**exponent\n",
    "    neurons_list.append([input_neurons] + [hidden_layer_neurons] * hidden_layer_count)\n",
    "    \n",
    "print(f'List of neurons: {neurons_list}')\n",
    "epochs = 100\n",
    "learning_rate = 0.01\n",
    "batch_size = 128\n",
    "characters_index = [2]\n",
    "\n",
    "n_iter_gibs = 1000\n",
    "n_imgs = 10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training with epochs = 100 \n",
      "neurons = [320, 16, 16, 16] \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training with epochs = 100 \n",
      "neurons = [320, 32, 32, 32] \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training with epochs = 100 \n",
      "neurons = [320, 64, 64, 64] \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training with epochs = 100 \n",
      "neurons = [320, 128, 128, 128] \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training with epochs = 100 \n",
      "neurons = [320, 256, 256, 256] \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training with epochs = 100 \n",
      "neurons = [320, 512, 512, 512] \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "all_generated_imgs = []\n",
    "\n",
    "image = lire_alpha_digit(binary_alpha_digs_path, characters_index)\n",
    "\n",
    "for index in range(len(neurons_list)):\n",
    "    \n",
    "    rbm = init_DBN(neurons_list[index])\n",
    "    rbm, loss = train_DBN(rbm, image, epochs, learning_rate, batch_size, verbose=False)\n",
    "\n",
    "    print('Training with epochs =' , epochs, '\\nneurons =', neurons_list[index], '\\nlearning_rate = ', learning_rate, '\\nbatch_size = ' , batch_size, '\\non characters : ', characters_index)\n",
    "    \n",
    "    generated_imgs = generer_image_DBN(rbm, n_imgs, n_iter_gibs)\n",
    "    \n",
    "    all_generated_imgs.append(generated_imgs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1080x864 with 60 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "n_rows = len(neurons_list)\n",
    "n_cols = n_imgs\n",
    "\n",
    "# Further adjust the figure size if needed\n",
    "fig, axs = plt.subplots(n_rows, n_cols, figsize=(15, 2 * n_rows))\n",
    "\n",
    "# Ensure axs is 2D for consistent indexing\n",
    "axs = axs.reshape(n_rows, n_cols) if n_rows > 1 else np.expand_dims(axs, 0)\n",
    "\n",
    "fig.patch.set_facecolor('black')\n",
    "\n",
    "# Set spacing between subplots and adjust margins\n",
    "plt.subplots_adjust(left=0.2, right=0.9, top=0.85, bottom=0.1, hspace=0.3, wspace=0.2)\n",
    "\n",
    "for i in range(n_rows):\n",
    "    for j in range(n_cols):\n",
    "        axs[i, j].imshow(all_generated_imgs[i][j].reshape(20, 16), cmap='gray')\n",
    "        axs[i, j].axis('off')\n",
    "\n",
    "        # Adjust the neuron label positioning and font\n",
    "        if j == 0:  # Only for the first column\n",
    "            axs[i, j].text(-0.5, 0.5, f'Neurons: {neurons_list[i][1]}', \n",
    "                           va='center', ha='left', rotation='vertical', \n",
    "                           fontsize=12, color=\"white\", fontweight='bold', transform=axs[i, j].transAxes)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2.2 The number of hidden layers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Neural network architectures: [[320, 128], [320, 128, 128], [320, 128, 128, 128], [320, 128, 128, 128, 128], [320, 128, 128, 128, 128, 128]]\n"
     ]
    }
   ],
   "source": [
    "input_layer_neurons = 320\n",
    "neurons_per_hidden_layer = 128\n",
    "min_hidden_layers = 1\n",
    "max_hidden_layers = 5\n",
    "\n",
    "neurons_list = []\n",
    "for layer_count in range(min_hidden_layers, max_hidden_layers + 1):\n",
    "    neurons_list.append([input_layer_neurons] + [neurons_per_hidden_layer] * layer_count)\n",
    "\n",
    "print(f'Neural network architectures: {neurons_list}')\n",
    "\n",
    "epochs = 100\n",
    "learning_rate = 0.01\n",
    "batch_size = 128\n",
    "characters_index = [2]\n",
    "\n",
    "n_iter_gibs = 1000\n",
    "n_imgs = 10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training with epochs = 100 \n",
      "neurons = [320, 128] \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training with epochs = 100 \n",
      "neurons = [320, 128, 128] \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training with epochs = 100 \n",
      "neurons = [320, 128, 128, 128] \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training with epochs = 100 \n",
      "neurons = [320, 128, 128, 128, 128] \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training with epochs = 100 \n",
      "neurons = [320, 128, 128, 128, 128, 128] \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAp4AAAEYCAYAAAD1d/bLAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/H5lhTAAAACXBIWXMAAAsTAAALEwEAmpwYAAAL30lEQVR4nO3d0Y7jNgwFUKfI///y9KnYznbtJhrmipLPeV1M4li0fCGA3MdxHF8HAAB82F+zLwAAgHsQPAEAiBA8AQCIEDwBAIgQPAEAiHimvujr67x5/vF4lH3e2WddfX+lkd/C6yrX0VrdT+U+NFKLaq6XxHuh85pXv5dXV10P8sifOfEEACBC8AQAIELwBAAgQvAEACBC8AQAIOJxFP9f7e92m1/9zV107Tzr6O5dqLzmLvvQrrU6ey1m10nlpJfqz9uh5mbXV2eJ9XXiCQBAhOAJAECE4AkAQITgCQBAhOAJAECE4AkAQET5OKUzI6Me7jLyYIfxFGfusoYjdl73SnevIXXyy4q10GE00h2tWCsjqkd/GacEAMA2BE8AACIETwAAIgRPAAAiBE8AACKeI39U3S2W6D6r7izkNdUdcmdrUrm+1V2CrOHd6RqpPaVrZ+oqRp7n2WubWr+7ds9X/rbKiT2pCQhXf5OoSSeeAABECJ4AAEQIngAARAieAABECJ4AAEQIngAARDyO4zjtq68c47HbSJDKsTqpcR98N3s00h3XMLWnVP5Npep9MLUPdeE90vv3zDB7T5kt9R4zTgkAgOUIngAARAieAABECJ4AAEQIngAARLToaq/slurcBd752n5qdq3Mrq9qO9TEO1bsJr0y8qyblPFL5279xLXt9jzwOStORnDiCQBAhOAJAECE4AkAQITgCQBAhOAJAEDEZVf7GR1387u4V7Hzb/u3FTsLZ1iho7obtfXLivtJ9SSMO75HqnWdmnCXZ92JJwAAEYInAAARgicAABGCJwAAEYInAAARgicAABFD45SqdR2xUj0G48zs3/lJu434UBOfsVudjOg64oXv3n1f2TP2tuJzO7LfVu7RTjwBAIgQPAEAiBA8AQCIEDwBAIgQPAEAiGjR1d6VbsS9VU5TGKkV634/1XuKGvqZxHNrb+CTVqwvJ54AAEQIngAARAieAABECJ4AAEQIngAARAieAABEGKd0oXpMQeX4Hj7HyBt+KjWK7Yya+xxryyq6Zg4nngAARAieAABECJ4AAEQIngAARAieAABEPGdfwG6uOh5nd5LxXWV3qrW9n1R3s9pam/WD75x4AgAQIXgCABAheAIAECF4AgAQIXgCABChq/2C/3edV5lmsL6uz7Ta+iV1L2b/f+zc00h9V9fqyOe9++w58QQAIELwBAAgQvAEACBC8AQAIELwBAAgQvAEACDicRzHknMjuo67uNt4kxUkasW67yu116ihz5j9ruiwrokROfTTdUScE08AACIETwAAIgRPAAAiBE8AACIETwAAIpbtan9XdWfj7K4wvtN5zCt0OPOKqzo5W8OuHcT0UzlloLJWr/6mkhNPAAAiBE8AACIETwAAIgRPAAAiBE8AACIETwAAIp6zL+CKETn3NHvUwxUjU/iHNV9D5eiaq89TD/zu3dobqaEV686JJwAAEYInAAARgicAABGCJwAAEYInAAARrbvaO9PZ+DkjHaXV38PaUhMxyKqeeFH5N94J/O7dWrnLJB8nngAARAieAABECJ4AAEQIngAARAieAABExLraZ3drjXx/dQelrsfXpGql83pU///SO5q9p9DLyPNsSsa5zvvjK1Lv790kngknngAARAieAABECJ4AAEQIngAARAieAABECJ4AAEQ8juOIzAjoPIqgcgRTyiojLa50vr8r2qEm/mTFsUmzR/uMWKF+RkbkzL6vV1LXvMLadtC5VkbMfiaMUwIAYCrBEwCACMETAIAIwRMAgAjBEwCAiPKu9q5dYZ27SUd061Jc8R7eRbdaeZfaylAnfXR+X3Wrk9n3o7OuUxOceAIAECF4AgAQIXgCABAheAIAECF4AgAQIXgCABAxNE6pw/iC1JiAd6XGYHQbaTEidT9m18SZ6mteuSaufm/XZz1lpE6q/2YFd6mH2Vapkw718O7e1eGdkHgvO/EEACBC8AQAIELwBAAgQvAEACBC8AQAIGKoqx0AAN7lxBMAgAjBEwCACMETAIAIwRMAgAjBEwCACMETAIAIwRMAgAjBEwCACMETAIAIwRMAgIhn9Qd+fWX+B87H4xH5Htb3bk2qrfWdrfnV2qb2rnepx9d497C6DjU8sne+y4knAAARgicAABGCJwAAEYInAAARgicAABGP4zjebqPq2v05Spfi5+xWK5XU3eeou3O71l3XNa+epLDr+nXXtb5SdLUDALAcwRMAgAjBEwCACMETAIAIwRMAgAjBEwCAiMtxSncfH3DFSIvvdqqVkbW9+v1q5TNWrLnq0Toj37Oy1JrPXif7yevO7lX1Pl5pZN07e/deO/EEACBC8AQAIELwBAAgQvAEACBC8AQAIKJFV/tOHV537Tic3ZE7+/uvVHZd7mqki7d6f5i9Hp1rOG3Fvf/M6muxqurJAImaTL17Us5+jxNPAAAiBE8AACIETwAAIgRPAAAiBE8AACIETwAAIp5X/1g5cmDks67+puuYAL5L3afKWh2pLfXwM9XPeuV6jOxps0doVY+SWcHs3zVSp7PrZAed72HiGmZnoRFOPAEAiBA8AQCIEDwBAIgQPAEAiBA8AQCIeBzHEWmJSnWmrtiBurPdumtnd1ivrrIDs/N9rdyHRuy633XYT2Z3Ea+wTkn25HNd9xQnngAARAieAABECJ4AAEQIngAARAieAABEXHa1d+5+1FnYy4qdhXfpsGZtKz5bP7XidJLUJIPV15Zz1VMbKuur8tqceAIAECF4AgAQIXgCABAheAIAECF4AgAQIXgCABBxOU5pRUbk8LvZNWFkyv+rHiNSeQ2V3z97DNxx3K+2qiXWsPPYqB3sdA9H9s7Z7yQnngAARAieAABECJ4AAEQIngAARAieAABEPGdfQAeprtUVO+b4bqQb0rqvYcV1WvGaV5CaPmD95rjLfe/ave/EEwCACMETAIAIwRMAgAjBEwCACMETAIAIwRMAgIjHcRyZuRHFKscEdB05wOtGxp+cra+RWMxgH+rDOCW62akmnXgCABAheAIAECF4AgAQIXgCABAheAIAEPGcfQHHMdbN+W5HcmW3++jnAfdmT7kf68qrUp3rsznxBAAgQvAEACBC8AQAIELwBAAgQvAEACCiRVd7dcc5fYx08VavbWICAvtK7TXqMW/knlsnPmWk7kY/byYnngAARAieAABECJ4AAEQIngAARAieAABECJ4AAES0GKc0IjGKp+sogpVUjyWxvuxKPfYxMgYOXrXTaKQRTjwBAIgQPAEAiBA8AQCIEDwBAIgQPAEAiGjR1V7Z3Zzodue/Rtbw3c+6slPHH3Mk9ojq56Hy83Z9hlL7yd3uK/WqJ7p0rT0nngAARAieAABECJ4AAEQIngAARAieAABECJ4AAEQ8juMonSFSORqpUuVoplHGbbymenwM/CM1Vk2d9jGyn9iDqDA723TlxBMAgAjBEwCACMETAIAIwRMAgAjBEwCAiGf1B450Cd7FSPfZ6p3w1d2hs+9H5ffrtl2D+7oG7xg+aXaH+tn3r/hOcOIJAECE4AkAQITgCQBAhOAJAECE4AkAQMTQ/9Wue7C263jkfnbrVhvp9r57HVXfm2418WnV96iyHjtPWZg9GeInqveM2VNYVpzssYPZ757Z6z67E96JJwAAEYInAAARgicAABGCJwAAEYInAAARgicAABHGKV3oPAqo2+iMytE2s+/tbrrVyp+kxpBV6rw/jFAn//9Zo5+3kxXq5DjW3FOuVI9NSji7ZieeAABECJ4AAEQIngAARAieAABECJ4AAEToah+U6opbpYNwhE74c5UdjKvX0E5r27lbevU6OTP7vu5mlTqx7vPpagcAYCrBEwCACMETAIAIwRMAgAjBEwCACMETAICI5+wLOI41R+RUjq65+p2rj8ipXsPKEUyV35HS+dpeMfv6U+OMVtzTVuD+8btUTYw805Xvnsrvn/0cOfEEACBC8AQAIELwBAAgQvAEACBC8AQAIOJxHIc2QQAAPs6JJwAAEYInAAARgicAABGCJwAAEYInAAARgicAABF/A4Q+0GZbKKweAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "all_generated_imgs = []\n",
    "\n",
    "image = lire_alpha_digit(binary_alpha_digs_path, characters_index)\n",
    "\n",
    "for index in range(len(neurons_list)):\n",
    "    \n",
    "    rbm_list = init_DBN(neurons_list[index])\n",
    "    rbm_list, loss = train_DBN(rbm_list, image, epochs, learning_rate, batch_size, verbose=False)\n",
    "\n",
    "    print('Training with epochs =' , epochs, '\\nneurons =', neurons_list[index], '\\nlearning_rate = ', learning_rate, '\\nbatch_size = ' , batch_size, '\\non characters : ', characters_index)\n",
    "    \n",
    "    generated_imgs = generer_image_DBN(rbm_list, n_imgs, n_iter_gibs)\n",
    "    \n",
    "    all_generated_imgs.append(generated_imgs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1080x720 with 50 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "n_rows = len(neurons_list)  \n",
    "n_cols = n_imgs\n",
    "\n",
    "# Further adjust the figure size if needed\n",
    "fig, axs = plt.subplots(n_rows, n_cols, figsize=(15, 2 * n_rows))\n",
    "\n",
    "# Ensure axs is 2D for consistent indexing\n",
    "axs = axs.reshape(n_rows, n_cols) if n_rows > 1 else np.expand_dims(axs, 0)\n",
    "\n",
    "fig.patch.set_facecolor('black')\n",
    "\n",
    "# Set spacing between subplots and adjust margins\n",
    "plt.subplots_adjust(left=0.2, right=0.9, top=0.85, bottom=0.1, hspace=0.3, wspace=0.2)\n",
    "\n",
    "for i in range(n_rows):\n",
    "    for j in range(n_cols):\n",
    "        axs[i, j].imshow(all_generated_imgs[i][j].reshape(20, 16), cmap='gray')\n",
    "        axs[i, j].axis('off')\n",
    "\n",
    "        # Adjust the neuron label positioning and font\n",
    "        if j == 0:  # Only for the first column\n",
    "            axs[i, j].text(-0.5, 0.5, f'Hidden layers: {len(neurons_list[i]) - 1}', \n",
    "                           va='center', ha='left', rotation='vertical', \n",
    "                           fontsize=10, color=\"white\", fontweight='bold', transform=axs[i, j].transAxes)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2.3 The number of characters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "n_neurons = [320] +  [128]*3\n",
    "epochs = 100\n",
    "learning_rate = 0.01\n",
    "batch_size = 128\n",
    "characters_list = [[2],  [2, 3], [2, 3, 4], [2, 3, 4, 5], [2, 3, 4, 5, 7], [2, 3, 4, 5, 7, 9]]\n",
    "\n",
    "n_iter_gibs = 1000\n",
    "n_imgs = 10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training with epochs = 100 \n",
      "neurons = [320, 64, 64, 64] \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training with epochs = 100 \n",
      "neurons = [320, 64, 64, 64] \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2, 3]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training with epochs = 100 \n",
      "neurons = [320, 64, 64, 64] \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2, 3, 4]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training with epochs = 100 \n",
      "neurons = [320, 64, 64, 64] \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2, 3, 4, 5]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training with epochs = 100 \n",
      "neurons = [320, 64, 64, 64] \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2, 3, 4, 5, 7]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training with epochs = 100 \n",
      "neurons = [320, 64, 64, 64] \n",
      "learning_rate =  0.01 \n",
      "batch_size =  128 \n",
      "on characters :  [2, 3, 4, 5, 7, 9]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x288 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "all_generated_imgs = []\n",
    "\n",
    "for index in range(len(characters_list)):\n",
    "    \n",
    "    image = lire_alpha_digit(binary_alpha_digs_path, characters_list[index])\n",
    "    \n",
    "    rbm = init_DBN(n_neurons)\n",
    "    rbm, loss = train_DBN(rbm, image, epochs, learning_rate, batch_size, verbose=False)\n",
    "\n",
    "    print('Training with epochs =' , epochs, '\\nneurons =', n_neurons, '\\nlearning_rate = ', learning_rate, '\\nbatch_size = ' , batch_size, '\\non characters : ', characters_list[index])\n",
    "\n",
    "    generated_imgs = generer_image_DBN(rbm, n_imgs, n_iter_gibs)\n",
    "    \n",
    "    all_generated_imgs.append(generated_imgs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1080x864 with 60 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "n_rows = len(characters_list)  \n",
    "n_cols = n_imgs\n",
    "\n",
    "# Further adjust the figure size if needed\n",
    "fig, axs = plt.subplots(n_rows, n_cols, figsize=(15, 2 * n_rows))\n",
    "\n",
    "# Ensure axs is 2D for consistent indexing\n",
    "axs = axs.reshape(n_rows, n_cols) if n_rows > 1 else np.expand_dims(axs, 0)\n",
    "\n",
    "fig.patch.set_facecolor('black')\n",
    "\n",
    "# Set spacing between subplots and adjust margins\n",
    "plt.subplots_adjust(left=0.2, right=0.9, top=0.85, bottom=0.1, hspace=0.3, wspace=0.2)\n",
    "\n",
    "for i in range(n_rows):\n",
    "    for j in range(n_cols):\n",
    "        axs[i, j].imshow(all_generated_imgs[i][j].reshape(20, 16), cmap='gray')\n",
    "        axs[i, j].axis('off')\n",
    "\n",
    "        # Adjust the neuron label positioning and font\n",
    "        if j == 0:  # Only for the first column\n",
    "            axs[i, j].text(-0.5, 0.5, f'{characters_list[i]}', \n",
    "                           va='center', ha='left', rotation='vertical', \n",
    "                           fontsize=12, color=\"white\", fontweight='bold', transform=axs[i, j].transAxes)\n",
    "\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "alte",
   "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": "3.9.12"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}