diff --git a/serie2/MNIST_binary_classifier_stud.ipynb b/serie2/MNIST_binary_classifier_stud.ipynb
index 2c976f6fded739ad1eaddb825816c4a8fe65e612..afc5755961f666461e6dae9859739cbf73bb197f 100644
--- a/serie2/MNIST_binary_classifier_stud.ipynb
+++ b/serie2/MNIST_binary_classifier_stud.ipynb
@@ -16,7 +16,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 1,
+   "execution_count": 2,
    "id": "educational-syndrome",
    "metadata": {},
    "outputs": [
@@ -41,7 +41,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 2,
+   "execution_count": 3,
    "id": "allied-flavor",
    "metadata": {},
    "outputs": [
@@ -76,7 +76,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 3,
+   "execution_count": 4,
    "id": "effective-anaheim",
    "metadata": {},
    "outputs": [
@@ -104,7 +104,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 4,
+   "execution_count": 5,
    "id": "stock-simpson",
    "metadata": {},
    "outputs": [],
@@ -145,7 +145,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 5,
+   "execution_count": 6,
    "id": "returning-relative",
    "metadata": {},
    "outputs": [
@@ -172,7 +172,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 6,
+   "execution_count": 7,
    "id": "qualified-charm",
    "metadata": {},
    "outputs": [
@@ -196,7 +196,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 7,
+   "execution_count": 8,
    "id": "signed-kansas",
    "metadata": {},
    "outputs": [
@@ -241,7 +241,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 116,
+   "execution_count": 34,
    "id": "removed-commons",
    "metadata": {},
    "outputs": [],
@@ -453,7 +453,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 123,
+   "execution_count": 40,
    "id": "colored-facility",
    "metadata": {},
    "outputs": [
@@ -461,17 +461,7 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "(1, 12136) (12136, 784) 784\n",
-      "(1, 12136) (12136, 784) 784\n",
-      "(1, 12136) (12136, 784) 784\n",
-      "(1, 12136) (12136, 784) 784\n",
-      "(1, 12136) (12136, 784) 784\n",
-      "(1, 12136) (12136, 784) 784\n",
-      "(1, 12136) (12136, 784) 784\n",
-      "(1, 12136) (12136, 784) 784\n",
-      "(1, 12136) (12136, 784) 784\n",
-      "(1, 12136) (12136, 784) 784\n",
-      "result after 10 epochs, train: cost 0.03605, error 0.08776 ; test: cost 0.03293, error 0.07482\n"
+      "result after 10 epochs, train: cost 0.01283, error 0.02620 ; test: cost 0.01404, error 0.02900\n"
      ]
     }
    ],
@@ -496,13 +486,24 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 127,
+   "execution_count": 41,
    "id": "lonely-quantity",
    "metadata": {},
    "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "[0. 1. 2. 3. 4. 5. 6. 7. 8. 9.]\n",
+      "[0.125      0.01536504 0.014986   0.01457691 0.01413949 0.01378105\n",
+      " 0.01344739 0.01318818 0.01299845 0.01282704]\n",
+      "[0.125      0.01556244 0.01537012 0.01520221 0.01504856 0.01491241\n",
+      " 0.01473657 0.01452109 0.01429011 0.01403994]\n"
+     ]
+    },
     {
      "data": {
-      "image/png": "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",
+      "image/png": "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",
       "text/plain": [
        "<Figure size 640x480 with 1 Axes>"
       ]
@@ -514,8 +515,11 @@
    "source": [
     "#analyse cost as function of epochs\n",
     "epochs = gradD.result_data[:,0]\n",
+    "print(epochs)\n",
     "train_costs = gradD.result_data[:,1]\n",
+    "print(train_costs)\n",
     "test_costs = gradD.result_data[:,3]\n",
+    "print(test_costs)\n",
     "\n",
     "plt.semilogy(epochs, train_costs, label=\"train\")\n",
     "plt.semilogy(epochs, test_costs, label=\"test\")\n",
@@ -531,7 +535,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 126,
+   "execution_count": 12,
    "id": "neither-moldova",
    "metadata": {},
    "outputs": [
@@ -540,15 +544,15 @@
      "output_type": "stream",
      "text": [
       "[0. 1. 2. 3. 4. 5. 6. 7. 8. 9.]\n",
-      "[0.47791694 0.16801252 0.12879038 0.1089321  0.09681938 0.09138102\n",
-      " 0.08948583 0.08857943 0.08833223 0.08775544]\n",
-      "[0.49208965 0.1723797  0.12755438 0.10415293 0.08536585 0.08075148\n",
-      " 0.07646671 0.07613711 0.07514832 0.07481872]\n"
+      "[0.47791694 0.0316414  0.030735   0.03024061 0.02916941 0.02859262\n",
+      " 0.02776862 0.02694463 0.02677983 0.02620303]\n",
+      "[0.49208965 0.0316414  0.0309822  0.0309822  0.03032301 0.03032301\n",
+      " 0.02999341 0.0306526  0.0306526  0.02900461]\n"
      ]
     },
     {
      "data": {
-      "image/png": "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",
+      "image/png": "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",
       "text/plain": [
        "<Figure size 640x480 with 1 Axes>"
       ]
@@ -582,7 +586,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 128,
+   "execution_count": 37,
    "id": "delayed-desire",
    "metadata": {},
    "outputs": [
@@ -590,7 +594,7 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "(227, 784)\n",
+      "(92, 784)\n",
       "ICI\n"
      ]
     },
@@ -601,14 +605,14 @@
      "traceback": [
       "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
       "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
-      "Cell \u001b[0;32mIn [128], line 28\u001b[0m\n\u001b[1;32m     26\u001b[0m \u001b[39mif\u001b[39;00m correct_labels\u001b[39m.\u001b[39mshape[\u001b[39m1\u001b[39m] \u001b[39m!=\u001b[39m rows\u001b[39m*\u001b[39mcols:\n\u001b[1;32m     27\u001b[0m     \u001b[39mprint\u001b[39m(\u001b[39m\"\u001b[39m\u001b[39mICI\u001b[39m\u001b[39m\"\u001b[39m)\n\u001b[0;32m---> 28\u001b[0m     \u001b[39mprint\u001b[39m(\u001b[39m-\u001b[39mnp\u001b[39m.\u001b[39mones((\u001b[39m1\u001b[39m,rows\u001b[39m*\u001b[39mcols\u001b[39m-\u001b[39mcorrect_labels\u001b[39m.\u001b[39mshape[\u001b[39m1\u001b[39m])))\n\u001b[1;32m     29\u001b[0m     correct_labels \u001b[39m=\u001b[39m np\u001b[39m.\u001b[39mappend(correct_labels, \u001b[39m-\u001b[39mnp\u001b[39m.\u001b[39mones((\u001b[39m1\u001b[39m,rows\u001b[39m*\u001b[39mcols\u001b[39m-\u001b[39mcorrect_labels\u001b[39m.\u001b[39mshape[\u001b[39m1\u001b[39m])),\u001b[39m1\u001b[39m)\n\u001b[1;32m     31\u001b[0m correct_labels \u001b[39m=\u001b[39m correct_labels\u001b[39m.\u001b[39mreshape(cols,rows)\u001b[39m.\u001b[39mT\n",
+      "Cell \u001b[0;32mIn [37], line 28\u001b[0m\n\u001b[1;32m     26\u001b[0m \u001b[39mif\u001b[39;00m correct_labels\u001b[39m.\u001b[39mshape[\u001b[39m1\u001b[39m] \u001b[39m!=\u001b[39m rows\u001b[39m*\u001b[39mcols:\n\u001b[1;32m     27\u001b[0m     \u001b[39mprint\u001b[39m(\u001b[39m\"\u001b[39m\u001b[39mICI\u001b[39m\u001b[39m\"\u001b[39m)\n\u001b[0;32m---> 28\u001b[0m     \u001b[39mprint\u001b[39m(\u001b[39m-\u001b[39mnp\u001b[39m.\u001b[39mones((\u001b[39m1\u001b[39m,rows\u001b[39m*\u001b[39mcols\u001b[39m-\u001b[39mcorrect_labels\u001b[39m.\u001b[39mshape[\u001b[39m1\u001b[39m])))\n\u001b[1;32m     29\u001b[0m     correct_labels \u001b[39m=\u001b[39m np\u001b[39m.\u001b[39mappend(correct_labels, \u001b[39m-\u001b[39mnp\u001b[39m.\u001b[39mones((\u001b[39m1\u001b[39m,rows\u001b[39m*\u001b[39mcols\u001b[39m-\u001b[39mcorrect_labels\u001b[39m.\u001b[39mshape[\u001b[39m1\u001b[39m])),\u001b[39m1\u001b[39m)\n\u001b[1;32m     31\u001b[0m correct_labels \u001b[39m=\u001b[39m correct_labels\u001b[39m.\u001b[39mreshape(cols,rows)\u001b[39m.\u001b[39mT\n",
       "File \u001b[0;32m~/.local/lib/python3.10/site-packages/numpy/core/numeric.py:204\u001b[0m, in \u001b[0;36mones\u001b[0;34m(shape, dtype, order, like)\u001b[0m\n\u001b[1;32m    201\u001b[0m \u001b[39mif\u001b[39;00m like \u001b[39mis\u001b[39;00m \u001b[39mnot\u001b[39;00m \u001b[39mNone\u001b[39;00m:\n\u001b[1;32m    202\u001b[0m     \u001b[39mreturn\u001b[39;00m _ones_with_like(shape, dtype\u001b[39m=\u001b[39mdtype, order\u001b[39m=\u001b[39morder, like\u001b[39m=\u001b[39mlike)\n\u001b[0;32m--> 204\u001b[0m a \u001b[39m=\u001b[39m empty(shape, dtype, order)\n\u001b[1;32m    205\u001b[0m multiarray\u001b[39m.\u001b[39mcopyto(a, \u001b[39m1\u001b[39m, casting\u001b[39m=\u001b[39m\u001b[39m'\u001b[39m\u001b[39munsafe\u001b[39m\u001b[39m'\u001b[39m)\n\u001b[1;32m    206\u001b[0m \u001b[39mreturn\u001b[39;00m a\n",
       "\u001b[0;31mValueError\u001b[0m: negative dimensions are not allowed"
      ]
     },
     {
      "data": {
-      "image/png": "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",
+      "image/png": "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",
       "text/plain": [
        "<Figure size 800x800 with 1 Axes>"
       ]
@@ -686,7 +690,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 129,
+   "execution_count": 42,
    "id": "another-setting",
    "metadata": {},
    "outputs": [],
@@ -725,7 +729,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 132,
+   "execution_count": 43,
    "id": "decreased-candidate",
    "metadata": {},
    "outputs": [
@@ -733,24 +737,23 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "(1, 5) (5, 3) 3\n",
-      "[[-0.09772056]\n",
-      " [-0.04798146]\n",
-      " [ 0.20071404]]\n",
+      "[[0.01439665]\n",
+      " [0.01451826]\n",
+      " [0.01512632]]\n",
       "-0.025\n"
      ]
     },
     {
      "ename": "AssertionError",
-     "evalue": "\nArrays are not almost equal to 8 decimals\n\nMismatched elements: 3 / 3 (100%)\nMax absolute difference: 0.17849182\nMax relative difference: 8.03213277\n x: array([[-0.09772056],\n       [-0.04798146],\n       [ 0.20071404]])\n y: array([[-0.01111111],\n       [-0.00555556],\n       [ 0.02222222]])",
+     "evalue": "\nArrays are not almost equal to 8 decimals\n\nMismatched elements: 3 / 3 (100%)\nMax absolute difference: 0.02550776\nMax relative difference: 3.61328441\n x: array([[0.01439665],\n       [0.01451826],\n       [0.01512632]])\n y: array([[-0.01111111],\n       [-0.00555556],\n       [ 0.02222222]])",
      "output_type": "error",
      "traceback": [
       "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
       "\u001b[0;31mAssertionError\u001b[0m                            Traceback (most recent call last)",
-      "Cell \u001b[0;32mIn [132], line 26\u001b[0m\n\u001b[1;32m     18\u001b[0m grad_w_exp \u001b[39m=\u001b[39m np\u001b[39m.\u001b[39marray([[\u001b[39m-\u001b[39m\u001b[39m0.01111111\u001b[39m],\n\u001b[1;32m     19\u001b[0m        [\u001b[39m-\u001b[39m\u001b[39m0.00555556\u001b[39m],\n\u001b[1;32m     20\u001b[0m        [ \u001b[39m0.02222222\u001b[39m]])\n\u001b[1;32m     22\u001b[0m grad_b_exp \u001b[39m=\u001b[39m np\u001b[39m.\u001b[39marray([\u001b[39m-\u001b[39m\u001b[39m0.025\u001b[39m])\n\u001b[0;32m---> 26\u001b[0m np\u001b[39m.\u001b[39mtesting\u001b[39m.\u001b[39massert_array_almost_equal(grad_w_pred\u001b[39m/\u001b[39m\u001b[39m2\u001b[39m,grad_w_exp,decimal\u001b[39m=\u001b[39m\u001b[39m8\u001b[39m)\n\u001b[1;32m     27\u001b[0m np\u001b[39m.\u001b[39mtesting\u001b[39m.\u001b[39massert_array_almost_equal(grad_b_pred,grad_b_exp,decimal\u001b[39m=\u001b[39m\u001b[39m8\u001b[39m)\n",
+      "Cell \u001b[0;32mIn [43], line 26\u001b[0m\n\u001b[1;32m     18\u001b[0m grad_w_exp \u001b[39m=\u001b[39m np\u001b[39m.\u001b[39marray([[\u001b[39m-\u001b[39m\u001b[39m0.01111111\u001b[39m],\n\u001b[1;32m     19\u001b[0m        [\u001b[39m-\u001b[39m\u001b[39m0.00555556\u001b[39m],\n\u001b[1;32m     20\u001b[0m        [ \u001b[39m0.02222222\u001b[39m]])\n\u001b[1;32m     22\u001b[0m grad_b_exp \u001b[39m=\u001b[39m np\u001b[39m.\u001b[39marray([\u001b[39m-\u001b[39m\u001b[39m0.025\u001b[39m])\n\u001b[0;32m---> 26\u001b[0m np\u001b[39m.\u001b[39mtesting\u001b[39m.\u001b[39massert_array_almost_equal(grad_w_pred,grad_w_exp,decimal\u001b[39m=\u001b[39m\u001b[39m8\u001b[39m)\n\u001b[1;32m     27\u001b[0m np\u001b[39m.\u001b[39mtesting\u001b[39m.\u001b[39massert_array_almost_equal(grad_b_pred,grad_b_exp,decimal\u001b[39m=\u001b[39m\u001b[39m8\u001b[39m)\n",
       "    \u001b[0;31m[... skipping hidden 1 frame]\u001b[0m\n",
       "File \u001b[0;32m~/.local/lib/python3.10/site-packages/numpy/testing/_private/utils.py:844\u001b[0m, in \u001b[0;36massert_array_compare\u001b[0;34m(comparison, x, y, err_msg, verbose, header, precision, equal_nan, equal_inf)\u001b[0m\n\u001b[1;32m    840\u001b[0m         err_msg \u001b[39m+\u001b[39m\u001b[39m=\u001b[39m \u001b[39m'\u001b[39m\u001b[39m\\n\u001b[39;00m\u001b[39m'\u001b[39m \u001b[39m+\u001b[39m \u001b[39m'\u001b[39m\u001b[39m\\n\u001b[39;00m\u001b[39m'\u001b[39m\u001b[39m.\u001b[39mjoin(remarks)\n\u001b[1;32m    841\u001b[0m         msg \u001b[39m=\u001b[39m build_err_msg([ox, oy], err_msg,\n\u001b[1;32m    842\u001b[0m                             verbose\u001b[39m=\u001b[39mverbose, header\u001b[39m=\u001b[39mheader,\n\u001b[1;32m    843\u001b[0m                             names\u001b[39m=\u001b[39m(\u001b[39m'\u001b[39m\u001b[39mx\u001b[39m\u001b[39m'\u001b[39m, \u001b[39m'\u001b[39m\u001b[39my\u001b[39m\u001b[39m'\u001b[39m), precision\u001b[39m=\u001b[39mprecision)\n\u001b[0;32m--> 844\u001b[0m         \u001b[39mraise\u001b[39;00m \u001b[39mAssertionError\u001b[39;00m(msg)\n\u001b[1;32m    845\u001b[0m \u001b[39mexcept\u001b[39;00m \u001b[39mValueError\u001b[39;00m:\n\u001b[1;32m    846\u001b[0m     \u001b[39mimport\u001b[39;00m \u001b[39mtraceback\u001b[39;00m\n",
-      "\u001b[0;31mAssertionError\u001b[0m: \nArrays are not almost equal to 8 decimals\n\nMismatched elements: 3 / 3 (100%)\nMax absolute difference: 0.17849182\nMax relative difference: 8.03213277\n x: array([[-0.09772056],\n       [-0.04798146],\n       [ 0.20071404]])\n y: array([[-0.01111111],\n       [-0.00555556],\n       [ 0.02222222]])"
+      "\u001b[0;31mAssertionError\u001b[0m: \nArrays are not almost equal to 8 decimals\n\nMismatched elements: 3 / 3 (100%)\nMax absolute difference: 0.02550776\nMax relative difference: 3.61328441\n x: array([[0.01439665],\n       [0.01451826],\n       [0.01512632]])\n y: array([[-0.01111111],\n       [-0.00555556],\n       [ 0.02222222]])"
      ]
     }
    ],
@@ -769,7 +772,7 @@
     "\n",
     "grad_w_pred, grad_b_pred = gradDummy.grad_cost()\n",
     "\n",
-    "print(grad_w_pred/2)\n",
+    "print(grad_w_pred)\n",
     "print(grad_b_pred)\n",
     "\n",
     "grad_w_exp = np.array([[-0.01111111],\n",
@@ -795,21 +798,29 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 29,
+   "execution_count": 47,
    "id": "hungry-electron",
    "metadata": {},
    "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "[[0.03455195 0.03484382 0.03630316]]\n",
+      "-0.1\n"
+     ]
+    },
     {
      "ename": "AssertionError",
-     "evalue": "\nArrays are not almost equal to 8 decimals\n\n(shapes (1, 784), (3, 1) mismatch)\n x: array([[0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n        0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n        0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,...\n y: array([[-0.04444444],\n       [-0.02222222],\n       [ 0.08888889]])",
+     "evalue": "\nArrays are not almost equal to 8 decimals\n\n(shapes (1, 3), (3, 1) mismatch)\n x: array([[0.03455195, 0.03484382, 0.03630316]])\n y: array([[-0.04444444],\n       [-0.02222222],\n       [ 0.08888889]])",
      "output_type": "error",
      "traceback": [
       "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
       "\u001b[0;31mAssertionError\u001b[0m                            Traceback (most recent call last)",
-      "Cell \u001b[0;32mIn [29], line 21\u001b[0m\n\u001b[1;32m     15\u001b[0m grad_w_exp \u001b[39m=\u001b[39m np\u001b[39m.\u001b[39marray([[\u001b[39m-\u001b[39m\u001b[39m0.04444444\u001b[39m],\n\u001b[1;32m     16\u001b[0m        [\u001b[39m-\u001b[39m\u001b[39m0.02222222\u001b[39m],\n\u001b[1;32m     17\u001b[0m        [ \u001b[39m0.08888889\u001b[39m]])\n\u001b[1;32m     19\u001b[0m grad_b_exp \u001b[39m=\u001b[39m np\u001b[39m.\u001b[39marray([\u001b[39m-\u001b[39m\u001b[39m0.1\u001b[39m])\n\u001b[0;32m---> 21\u001b[0m np\u001b[39m.\u001b[39mtesting\u001b[39m.\u001b[39massert_array_almost_equal(grad_w_pred,grad_w_exp,decimal\u001b[39m=\u001b[39m\u001b[39m8\u001b[39m)\n\u001b[1;32m     22\u001b[0m np\u001b[39m.\u001b[39mtesting\u001b[39m.\u001b[39massert_array_almost_equal(grad_b_pred,grad_b_exp,decimal\u001b[39m=\u001b[39m\u001b[39m8\u001b[39m)\n",
+      "Cell \u001b[0;32mIn [47], line 24\u001b[0m\n\u001b[1;32m     18\u001b[0m grad_w_exp \u001b[39m=\u001b[39m np\u001b[39m.\u001b[39marray([[\u001b[39m-\u001b[39m\u001b[39m0.04444444\u001b[39m],\n\u001b[1;32m     19\u001b[0m        [\u001b[39m-\u001b[39m\u001b[39m0.02222222\u001b[39m],\n\u001b[1;32m     20\u001b[0m        [ \u001b[39m0.08888889\u001b[39m]])\n\u001b[1;32m     22\u001b[0m grad_b_exp \u001b[39m=\u001b[39m np\u001b[39m.\u001b[39marray([\u001b[39m-\u001b[39m\u001b[39m0.1\u001b[39m])\n\u001b[0;32m---> 24\u001b[0m np\u001b[39m.\u001b[39mtesting\u001b[39m.\u001b[39massert_array_almost_equal(grad_w_pred,grad_w_exp,decimal\u001b[39m=\u001b[39m\u001b[39m8\u001b[39m)\n\u001b[1;32m     25\u001b[0m np\u001b[39m.\u001b[39mtesting\u001b[39m.\u001b[39massert_array_almost_equal(grad_b_pred,grad_b_exp,decimal\u001b[39m=\u001b[39m\u001b[39m8\u001b[39m)\n",
       "    \u001b[0;31m[... skipping hidden 1 frame]\u001b[0m\n",
       "File \u001b[0;32m~/.local/lib/python3.10/site-packages/numpy/testing/_private/utils.py:763\u001b[0m, in \u001b[0;36massert_array_compare\u001b[0;34m(comparison, x, y, err_msg, verbose, header, precision, equal_nan, equal_inf)\u001b[0m\n\u001b[1;32m    757\u001b[0m \u001b[39mif\u001b[39;00m \u001b[39mnot\u001b[39;00m cond:\n\u001b[1;32m    758\u001b[0m     msg \u001b[39m=\u001b[39m build_err_msg([x, y],\n\u001b[1;32m    759\u001b[0m                         err_msg\n\u001b[1;32m    760\u001b[0m                         \u001b[39m+\u001b[39m \u001b[39mf\u001b[39m\u001b[39m'\u001b[39m\u001b[39m\\n\u001b[39;00m\u001b[39m(shapes \u001b[39m\u001b[39m{\u001b[39;00mx\u001b[39m.\u001b[39mshape\u001b[39m}\u001b[39;00m\u001b[39m, \u001b[39m\u001b[39m{\u001b[39;00my\u001b[39m.\u001b[39mshape\u001b[39m}\u001b[39;00m\u001b[39m mismatch)\u001b[39m\u001b[39m'\u001b[39m,\n\u001b[1;32m    761\u001b[0m                         verbose\u001b[39m=\u001b[39mverbose, header\u001b[39m=\u001b[39mheader,\n\u001b[1;32m    762\u001b[0m                         names\u001b[39m=\u001b[39m(\u001b[39m'\u001b[39m\u001b[39mx\u001b[39m\u001b[39m'\u001b[39m, \u001b[39m'\u001b[39m\u001b[39my\u001b[39m\u001b[39m'\u001b[39m), precision\u001b[39m=\u001b[39mprecision)\n\u001b[0;32m--> 763\u001b[0m     \u001b[39mraise\u001b[39;00m \u001b[39mAssertionError\u001b[39;00m(msg)\n\u001b[1;32m    765\u001b[0m flagged \u001b[39m=\u001b[39m bool_(\u001b[39mFalse\u001b[39;00m)\n\u001b[1;32m    766\u001b[0m \u001b[39mif\u001b[39;00m isnumber(x) \u001b[39mand\u001b[39;00m isnumber(y):\n",
-      "\u001b[0;31mAssertionError\u001b[0m: \nArrays are not almost equal to 8 decimals\n\n(shapes (1, 784), (3, 1) mismatch)\n x: array([[0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n        0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n        0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,...\n y: array([[-0.04444444],\n       [-0.02222222],\n       [ 0.08888889]])"
+      "\u001b[0;31mAssertionError\u001b[0m: \nArrays are not almost equal to 8 decimals\n\n(shapes (1, 3), (3, 1) mismatch)\n x: array([[0.03455195, 0.03484382, 0.03630316]])\n y: array([[-0.04444444],\n       [-0.02222222],\n       [ 0.08888889]])"
      ]
     }
    ],
@@ -828,6 +839,9 @@
     "\n",
     "grad_w_pred, grad_b_pred = gradDummy.grad_cost()\n",
     "\n",
+    "print(grad_w_pred)\n",
+    "print(grad_b_pred)\n",
+    "\n",
     "grad_w_exp = np.array([[-0.04444444],\n",
     "       [-0.02222222],\n",
     "       [ 0.08888889]])\n",
@@ -848,7 +862,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 44,
    "id": "identical-worthy",
    "metadata": {},
    "outputs": [],
@@ -883,7 +897,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 45,
    "id": "chemical-nothing",
    "metadata": {},
    "outputs": [],
@@ -919,7 +933,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 14,
    "id": "90f77759",
    "metadata": {},
    "outputs": [
@@ -969,7 +983,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 15,
    "id": "5e873ea8",
    "metadata": {},
    "outputs": [
@@ -995,9 +1009,48 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 25,
    "id": "85bb33b8",
    "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "[<matplotlib.lines.Line2D at 0x7f08b8782e90>]"
+      ]
+     },
+     "execution_count": 25,
+     "metadata": {},
+     "output_type": "execute_result"
+    },
+    {
+     "data": {
+      "image/png": "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",
+      "text/plain": [
+       "<Figure size 640x480 with 1 Axes>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "def c1(z):\n",
+    "  return (sigmo(z)-1)**2\n",
+    "\n",
+    "\n",
+    "plt.plot(x, c1(x))\n",
+    "plt.grid()\n",
+    "\n",
+    "p1 = [-4, c1(-4)]\n",
+    "p2 = [4, c1(4)]\n",
+    "plt.plot([p1[0], p2[0]], [p1[1], p2[1]])\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
    "outputs": [],
    "source": []
   }
diff --git a/serie2/TSM_DeLear - TP2.2.pdf b/serie2/TSM_DeLear - TP2.2.pdf
new file mode 100644
index 0000000000000000000000000000000000000000..ad087c68263764752e3ebf253f30d0a22af87439
Binary files /dev/null and b/serie2/TSM_DeLear - TP2.2.pdf differ
diff --git a/serie2/cz.png b/serie2/cz.png
new file mode 100644
index 0000000000000000000000000000000000000000..0ce93da1cc7aa67c2368398e25c4281959e376ef
Binary files /dev/null and b/serie2/cz.png differ
diff --git a/serie2/dsigmo.png b/serie2/dsigmo.png
new file mode 100644
index 0000000000000000000000000000000000000000..dc63bb768cc1dbc1391f0671b8a4cb683c047ab1
Binary files /dev/null and b/serie2/dsigmo.png differ
diff --git a/serie2/nonconvex.png b/serie2/nonconvex.png
new file mode 100644
index 0000000000000000000000000000000000000000..d472145ab1fc9bd10f03c9c94083cf4b54e912c5
Binary files /dev/null and b/serie2/nonconvex.png differ
diff --git a/serie2/outputCost.png b/serie2/outputCost.png
new file mode 100644
index 0000000000000000000000000000000000000000..8ed49c332a2bb2a52f8fa2c6f1f8b9199b64af6d
Binary files /dev/null and b/serie2/outputCost.png differ
diff --git a/serie2/outputError.png b/serie2/outputError.png
new file mode 100644
index 0000000000000000000000000000000000000000..53b2ead6ef65970771882bf3dd61261fbd4eba1a
Binary files /dev/null and b/serie2/outputError.png differ
diff --git a/serie2/rapport.md b/serie2/rapport.md
index 99aba2fe99b10aee55c64859a994580a4761d680..5f498f48c7f0367ea1754f994160fa7498c7751a 100644
--- a/serie2/rapport.md
+++ b/serie2/rapport.md
@@ -1,6 +1,120 @@
-# Série 2
+# Series 2
 Simon Cirilli - Kiady Arintsoa - Teo Colomboretto
-## Partie 2
+## Section 2
+As for part 2 of this work, we have largely solved everything on paper. This work is present in the file "TSM_DeLear - TP2.2.pdf". What is missing in the pdf is therefore present in this part. 
 
+### c)
+The calculation of the asymptotes are for -inf and +inf :
+$$
+\lim_{z \to -\infty} -ln(\frac{1}{1+e^{+z}})= \lim_{z \to -\infty} -ln(->0) = - (-\infty) = +\infty 
+$$
+
+$$
+\lim_{z \to +\infty} -ln(\frac{1}{1+e^{+z}})= \lim_{z \to +\infty} -ln(->1) = - (0) = 0
+$$
+
+The graph of this function is :
+![softplus](cz.png)
+
+### d)
+```python
+def sigmo(z):
+    return 1/(1+np.exp(-z))
+
+def derivat_sigmo(z):
+    return np.exp(-z)/((1+np.exp(-z))**2)
+
+x = np.linspace(-10,10,100)
+y = sigmo(x)
+plt.plot(x,y)
+plt.grid()
+plt.show()
+
+
+x = np.linspace(-10,10,100)
+y = derivat_sigmo(x)
+plt.plot(x,y)
+plt.grid()
+plt.show()
+
+```
+![sigmo](sigmo.png)
+![derivat_sigmo](dsigmo.png)
+
+### f)
+```python
+def c1(z):
+  return (sigmo(z)-1)**2
+
+plt.plot(x, c1(x))
+plt.grid()
+
+p1 = [-4, c1(-4)]
+p2 = [4, c1(4)]
+plt.plot([p1[0], p2[0]], [p1[1], p2[1]])
+```
+![c1](nonconvex.png)
+
+We notice indeed that the function is not convex because by opposing it to a straight line, we obtain a curve which is not like a "U" but more like a "W".
+
+<!-- Explain in which situations (initial settings) optimising c1(x) with gradient descent may become difficult.  -->
+So in which situations (initial settings) optimising c1(x) with gradient descent may become difficult ? In the case where the initial point is very close to the minimum, the gradient descent will not be able to converge correctly because the gradient will be very small and the steps will be very small.
+
+## Section 3
+### Description of the code
+For this section we had to study the GradientDescent class and we had to implement some functions of it.
+
+The first part to implement is the data normalization function. In our case we decided to do a z-norm by calculating the mean and the standard deviation of each column of the data matrix. We then applied the following formula to each element of the matrix.
+
+The second part consisted in implementing the MSE cost function or the cross-entropy.
+
+The third part consisted in implementing the cost of the gradient according to the cost function used (either MSE or CE).
+
+Finally, the last part consisted in implementing the sigmoid function as well as the update function that performed an iteration of the gradient descent.
+
+### Result
+We have tested our code with the unit test but we have some few failures. The predict part and the cost func part seems to work nicely but not the grad cost part. We have the same value for the bias but not for the weights. Effectively we have :
+```python
+# MSE grad cost
+# calculated weights
+[[0.01439665]
+ [0.01451826]
+ [0.01512632]]
+-0.025
+
+# expected weights
+[[-0.01111111],
+ [-0.00555556],
+ [ 0.02222222]]
+-0.025
+
+
+# CE grad cost
+# calculated weights
+[[0.03455195 0.03484382 0.03630316]]
+-0.1
+
+
+# expected weights
+[[-0.04444444],
+ [-0.02222222],
+ [ 0.08888889]]
+-0.1
+```
+But we see that the gradient descent with precedent graphs is working well. We have a good convergence and the cost is decreasing with 10 epochs.
+![cost](outputCost.png)
+
+![error](outputError.png)
+
+With value for the cost decreasing :
+```python
+# train cost
+[0.125      0.01536504 0.014986   0.01457691 0.01413949 0.01378105
+ 0.01344739 0.01318818 0.01299845 0.01282704]
+# test cost
+[0.125      0.01556244 0.01537012 0.01520221 0.01504856 0.01491241
+ 0.01473657 0.01452109 0.01429011 0.01403994]
+```
+
+### Analyse the dependency on the learning rate
 
-## Partie 3
\ No newline at end of file
diff --git a/serie2/sigmo.png b/serie2/sigmo.png
new file mode 100644
index 0000000000000000000000000000000000000000..50a39578cb4aa8286c503a05fe7e9a973c082b4c
Binary files /dev/null and b/serie2/sigmo.png differ