diff --git a/app/function_approximation.py b/app/function_approximation.py
index d2cd790ac9e5759b79b754fe17c837109579ab04..aee561f11a318fe554d5802367a98a5ba800bab0 100644
--- a/app/function_approximation.py
+++ b/app/function_approximation.py
@@ -1,20 +1,44 @@
+import os
 import numpy as np
-import matplotlib.pyplot as plt
 import torch
-from src.misc import time_stamp, timeit
-from src import target_function
+import matplotlib.pyplot as plt
+import neural_networks_101.src as src
 
 
 def main() -> None:
-    print(time_stamp(), "Initialize main file")
-    with timeit("create x_train data ({:4.2f} s)"):
-        x_train = np.random.uniform(0, 1, (100000, 2))
-    with timeit("create y_train data ({:4.2f} s)"):
-        y_train = target_function.sin2d(x_train)
-
-    plt.figure()
-    plt.hexbin(x_train[:, 0], x_train[:, 1], y_train, gridsize=50)
-    plt.show()
+    # Get CPU or GPU device for training
+    device = "cuda" if torch.cuda.is_available() else "cpu"
+    device = torch.device(device)
+
+    # generate samples
+    print(src.misc.time_stamp(), "generate training and test data")
+    x_train = np.random.uniform(0, 1, (100000, 2))
+    x_test = np.random.uniform(0, 1, (10000, 2))
+    y_train = src.target_function.sin2d(x_train).reshape(-1, 1)
+    y_test = src.target_function.sin2d(x_test).reshape(-1, 1)
+
+    # define model, loss and optimization algorithm
+    model = src.approximation.NeuralNetwork(
+        x_train.shape[1], y_train.shape[1], width=1024).to(device)
+    optimizer = torch.optim.Adam(model.parameters(), lr=1e-02)
+    loss_function = torch.nn.MSELoss(reduction="mean")
+
+    n_epochs = 5
+    for epoch in range(n_epochs):
+        with src.misc.timeit("time: {:4.2f} s"):
+            src.approximation.train(
+                model, device, x_train, y_train, loss_function, optimizer, log_interval=100)
+        test_loss = src.approximation.test(
+            model, device, x_test, y_test, loss_function)
+        print(src.misc.time_stamp(), f"test set avg. loss: {test_loss}")
+
+    if not os.path.isdir("../img"):
+        os.makedirs("../img", exist_ok=True)
+    src.approximation.plot_function_approximation(
+        model, device, src.target_function.sin2d, figsize=(15, 5))
+    file_name = "../img/function_approximation.png"
+    plt.savefig(file_name, dpi=200)
+    print(src.misc.time_stamp(), f"save to: {file_name}")
 
 
 if __name__ == "__main__":
diff --git a/nbs/function_approximation.ipynb b/nbs/function_approximation.ipynb
index 660fb2a7b35e5cad61be5e693a5abbe3307bc7a5..408d50897448f61b2bb211fa6763066e23f2e9a8 100644
--- a/nbs/function_approximation.ipynb
+++ b/nbs/function_approximation.ipynb
@@ -5,8 +5,8 @@
    "execution_count": 1,
    "metadata": {
     "ExecuteTime": {
-     "end_time": "2022-06-27T06:51:44.140941Z",
-     "start_time": "2022-06-27T06:51:43.403583Z"
+     "end_time": "2022-07-07T07:55:37.379253Z",
+     "start_time": "2022-07-07T07:55:36.412460Z"
     }
    },
    "outputs": [],
@@ -26,8 +26,8 @@
    "execution_count": 2,
    "metadata": {
     "ExecuteTime": {
-     "end_time": "2022-06-27T06:51:44.157724Z",
-     "start_time": "2022-06-27T06:51:44.143028Z"
+     "end_time": "2022-07-07T07:55:37.402397Z",
+     "start_time": "2022-07-07T07:55:37.382145Z"
     }
    },
    "outputs": [],
@@ -42,8 +42,8 @@
    "execution_count": 3,
    "metadata": {
     "ExecuteTime": {
-     "end_time": "2022-06-27T06:51:44.184037Z",
-     "start_time": "2022-06-27T06:51:44.164227Z"
+     "end_time": "2022-07-07T07:55:37.432619Z",
+     "start_time": "2022-07-07T07:55:37.404579Z"
     }
    },
    "outputs": [
@@ -51,14 +51,14 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "[2022-06-27 08:51:44] generate training and test data\n"
+      "[2022-07-07 09:55:37] generate training and test data\n"
      ]
     }
    ],
    "source": [
     "# generate samples\n",
     "print(src.misc.time_stamp(), \"generate training and test data\")\n",
-    "x_train = np.random.uniform(0, 1, (10000, 2))\n",
+    "x_train = np.random.uniform(0, 1, (100000, 2))\n",
     "x_test = np.random.uniform(0, 1, (1000, 2))\n",
     "y_train = src.target_function.sin2d(x_train).reshape(-1, 1)\n",
     "y_test = src.target_function.sin2d(x_test).reshape(-1, 1)"
@@ -69,8 +69,8 @@
    "execution_count": 4,
    "metadata": {
     "ExecuteTime": {
-     "end_time": "2022-06-27T06:51:44.209802Z",
-     "start_time": "2022-06-27T06:51:44.185625Z"
+     "end_time": "2022-07-07T07:55:37.467293Z",
+     "start_time": "2022-07-07T07:55:37.435384Z"
     }
    },
    "outputs": [],
@@ -86,8 +86,8 @@
    "execution_count": 5,
    "metadata": {
     "ExecuteTime": {
-     "end_time": "2022-06-27T06:52:01.055372Z",
-     "start_time": "2022-06-27T06:51:44.211488Z"
+     "end_time": "2022-07-07T08:00:07.997406Z",
+     "start_time": "2022-07-07T07:55:37.470695Z"
     }
    },
    "outputs": [
@@ -95,22 +95,78 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "progress:  0.0 % -- loss: 0.2509300410747528\n",
-      "progress: 63.7 % -- loss: 0.22826001048088074\n",
-      "[2022-06-27 08:51:48] time: 3.93 s\n",
-      "[2022-06-27 08:51:48] test set avg. loss: 0.015558036975562572\n",
-      "progress:  0.0 % -- loss: 0.30233126878738403\n",
-      "progress: 63.7 % -- loss: 0.20259657502174377\n",
-      "[2022-06-27 08:51:52] time: 3.79 s\n",
-      "[2022-06-27 08:51:52] test set avg. loss: 0.0056471554562449455\n",
-      "progress:  0.0 % -- loss: 0.0772915855050087\n",
-      "progress: 63.7 % -- loss: 0.07037431001663208\n",
-      "[2022-06-27 08:51:56] time: 3.75 s\n",
-      "[2022-06-27 08:51:56] test set avg. loss: 0.004849676508456469\n",
-      "progress:  0.0 % -- loss: 0.0791594386100769\n",
-      "progress: 63.7 % -- loss: 0.025842074304819107\n",
-      "[2022-06-27 08:52:00] time: 4.19 s\n",
-      "[2022-06-27 08:52:01] test set avg. loss: 0.0014430878218263388\n"
+      "progress:  0.0 % -- loss: 0.2427394688129425\n",
+      "progress:  6.4 % -- loss: 0.1920534372329712\n",
+      "progress: 12.8 % -- loss: 0.2313966602087021\n",
+      "progress: 19.2 % -- loss: 0.2262098491191864\n",
+      "progress: 25.6 % -- loss: 0.18118232488632202\n",
+      "progress: 32.0 % -- loss: 0.21185600757598877\n",
+      "progress: 38.4 % -- loss: 0.07742904871702194\n",
+      "progress: 44.8 % -- loss: 0.09230078756809235\n",
+      "progress: 51.2 % -- loss: 0.04743719846010208\n",
+      "progress: 57.6 % -- loss: 0.0509481355547905\n",
+      "progress: 64.0 % -- loss: 0.01298436988145113\n",
+      "progress: 70.4 % -- loss: 0.02120228484272957\n",
+      "progress: 76.8 % -- loss: 0.013040924444794655\n",
+      "progress: 83.2 % -- loss: 0.014941082336008549\n",
+      "progress: 89.6 % -- loss: 0.007720808498561382\n",
+      "progress: 96.0 % -- loss: 0.02163536474108696\n",
+      "[2022-07-07 09:56:47] time: 70.12 s\n",
+      "[2022-07-07 09:56:48] test set avg. loss: 0.0009188529220409691\n",
+      "progress:  0.0 % -- loss: 0.013689007610082626\n",
+      "progress:  6.4 % -- loss: 0.004026013892143965\n",
+      "progress: 12.8 % -- loss: 0.0031869893427938223\n",
+      "progress: 19.2 % -- loss: 0.003966950345784426\n",
+      "progress: 25.6 % -- loss: 0.003397703170776367\n",
+      "progress: 32.0 % -- loss: 0.005538471508771181\n",
+      "progress: 38.4 % -- loss: 0.010127665475010872\n",
+      "progress: 44.8 % -- loss: 0.005245153792202473\n",
+      "progress: 51.2 % -- loss: 0.0057514808140695095\n",
+      "progress: 57.6 % -- loss: 0.017985671758651733\n",
+      "progress: 64.0 % -- loss: 0.014705226756632328\n",
+      "progress: 70.4 % -- loss: 0.0009815238881856203\n",
+      "progress: 76.8 % -- loss: 0.002165044192224741\n",
+      "progress: 83.2 % -- loss: 0.009639129042625427\n",
+      "progress: 89.6 % -- loss: 0.01951625384390354\n",
+      "progress: 96.0 % -- loss: 0.019376104697585106\n",
+      "[2022-07-07 09:58:06] time: 77.98 s\n",
+      "[2022-07-07 09:58:06] test set avg. loss: 0.0006797168753109872\n",
+      "progress:  0.0 % -- loss: 0.014172333292663097\n",
+      "progress:  6.4 % -- loss: 0.004388859495520592\n",
+      "progress: 12.8 % -- loss: 0.0015614626463502645\n",
+      "progress: 19.2 % -- loss: 0.002652204129844904\n",
+      "progress: 25.6 % -- loss: 0.002200125716626644\n",
+      "progress: 32.0 % -- loss: 0.044217299669981\n",
+      "progress: 38.4 % -- loss: 0.008509873412549496\n",
+      "progress: 44.8 % -- loss: 0.010874507017433643\n",
+      "progress: 51.2 % -- loss: 0.0013758536661043763\n",
+      "progress: 57.6 % -- loss: 0.0010635185753926635\n",
+      "progress: 64.0 % -- loss: 0.004772291984409094\n",
+      "progress: 70.4 % -- loss: 0.001246526138857007\n",
+      "progress: 76.8 % -- loss: 0.0018227449618279934\n",
+      "progress: 83.2 % -- loss: 0.003252860624343157\n",
+      "progress: 89.6 % -- loss: 0.005041578318923712\n",
+      "progress: 96.0 % -- loss: 0.0036862825509160757\n",
+      "[2022-07-07 09:59:00] time: 54.11 s\n",
+      "[2022-07-07 09:59:00] test set avg. loss: 0.0006144134094938636\n",
+      "progress:  0.0 % -- loss: 0.010075250640511513\n",
+      "progress:  6.4 % -- loss: 0.007143543567508459\n",
+      "progress: 12.8 % -- loss: 0.004148718900978565\n",
+      "progress: 19.2 % -- loss: 0.004274588543921709\n",
+      "progress: 25.6 % -- loss: 0.003818443277850747\n",
+      "progress: 32.0 % -- loss: 0.007201837375760078\n",
+      "progress: 38.4 % -- loss: 0.0036278916522860527\n",
+      "progress: 44.8 % -- loss: 0.0012342752888798714\n",
+      "progress: 51.2 % -- loss: 0.015445721335709095\n",
+      "progress: 57.6 % -- loss: 0.016958491876721382\n",
+      "progress: 64.0 % -- loss: 0.0056175063364207745\n",
+      "progress: 70.4 % -- loss: 0.005791833624243736\n",
+      "progress: 76.8 % -- loss: 0.0010696140816435218\n",
+      "progress: 83.2 % -- loss: 0.000853274657856673\n",
+      "progress: 89.6 % -- loss: 0.0004952493472956121\n",
+      "progress: 96.0 % -- loss: 0.0022959031630307436\n",
+      "[2022-07-07 10:00:07] time: 66.76 s\n",
+      "[2022-07-07 10:00:07] test set avg. loss: 9.420912101631984e-05\n"
      ]
     }
    ],
@@ -128,16 +184,16 @@
    "execution_count": 6,
    "metadata": {
     "ExecuteTime": {
-     "end_time": "2022-06-27T06:52:01.684805Z",
-     "start_time": "2022-06-27T06:52:01.061488Z"
+     "end_time": "2022-07-07T08:00:08.887327Z",
+     "start_time": "2022-07-07T08:00:08.004762Z"
     }
    },
    "outputs": [
     {
      "data": {
-      "image/png": "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\n",
+      "image/png": "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\n",
       "text/plain": [
-       "<Figure size 432x288 with 3 Axes>"
+       "<Figure size 1080x360 with 6 Axes>"
       ]
      },
      "metadata": {
@@ -147,13 +203,7 @@
     }
    ],
    "source": [
-    "fig, axes = plt.subplot_mosaic([[\"true\", \"approx\", \"error\"]])\n",
-    "x = np.linspace(0, 1, 50)\n",
-    "X, Y = np.meshgrid(x,x)\n",
-    "xx = np.concatenate([X.reshape(-1, 1), Y.reshape(-1, 1)], axis=1)\n",
-    "axes[\"true\"].contourf(x, x, src.target_function.sin2d(xx).reshape(x.size, -1))\n",
-    "axes[\"approx\"].contourf(x, x, src.approximation.evaluate(model, device, xx).reshape(x.size, -1))\n",
-    "im = axes[\"error\"].contourf(x, x, src.target_function.sin2d(xx).reshape(x.size, -1)-src.approximation.evaluate(model, device, xx).reshape(x.size, -1))"
+    "src.approximation.plot_function_approximation(model, device, src.target_function.sin2d, figsize=(15, 5))"
    ]
   },
   {
diff --git a/src/approximation.py b/src/approximation.py
index d88b7488a46a132ca57857aa61375551f04fe7a3..d6d7918933d78964723845a72a98de45cb7a6b04 100644
--- a/src/approximation.py
+++ b/src/approximation.py
@@ -168,3 +168,29 @@ def evaluate(model: NeuralNetwork,
     model.eval()
     xs = torch.from_numpy(xs).type(torch.float32).to(device)
     return model(xs).detach().numpy()
+
+
+def plot_function_approximation(model, device, target, **kwargs):
+    """ Plot function approximation error.
+
+    Parameters
+    ----------
+    model : NeuralNetwork
+        Neural network model.
+    device : torch.device
+        Hardware to train the model on.
+    target : Callable
+        Target function.
+    """
+    fig, axes = plt.subplot_mosaic([["true", "approx", "error"]], **kwargs)
+    x = np.linspace(0, 1, 50)
+    X, Y = np.meshgrid(x, x)
+    xx = np.concatenate([X.reshape(-1, 1), Y.reshape(-1, 1)], axis=1)
+    f_val = target(xx).reshape(x.size, -1)
+    f_nn = evaluate(model, device, xx).reshape(x.size, -1)
+    im = axes["true"].contourf(x, x, f_val)
+    plt.colorbar(im, ax=axes["true"])
+    im = axes["approx"].contourf(x, x, f_nn)
+    plt.colorbar(im, ax=axes["approx"])
+    im = axes["error"].contourf(x, x, f_val-f_nn)
+    plt.colorbar(im, ax=axes["error"])