Removed use of torch_lr_finder

- Since the environment will not have torch-lr-finder I've removed the
import where it was unused and commented it out otherwise.

notebooks/LRP-L96.ipynb

@@ -6,19 +6,11 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "%matplotlib inline\n",
-    "import math\n",
-    "\n",
+    "# import math\n",
     "import matplotlib.pyplot as plt\n",
     "import numpy as np\n",
     "import torch\n",
-    "import torch.nn.functional as F\n",
-    "import torch.utils.data as Data\n",
-    "import torchvision\n",
-    "from sklearn.metrics import r2_score\n",
-    "from torch import nn, optim\n",
-    "from torch.autograd import Variable\n",
-    "from torch_lr_finder import LRFinder\n",
+    "from torch import nn\n",
     "\n",
     "np.random.seed(14)  # For reproducibility\n",
     "torch.manual_seed(14)  # For reproducibility"

notebooks/Neural-Network-Advection-FwdEuler.ipynb

@@ -13,9 +13,6 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "%matplotlib inline\n",
-    "import math\n",
-    "\n",
     "import matplotlib.pyplot as plt\n",
     "import numpy as np\n",
     "from IPython.display import HTML\n",
@@ -709,7 +706,6 @@
     "from sklearn.metrics import r2_score\n",
     "from torch import nn, optim\n",
     "from torch.autograd import Variable\n",
-    "from torch_lr_finder import LRFinder\n",
     "\n",
     "np.random.seed(14)  # For reproducibility\n",
     "torch.manual_seed(14)  # For reproducibility"

notebooks/Neural-Network-Advection.ipynb

@@ -13,9 +13,6 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "%matplotlib inline\n",
-    "import math\n",
-    "\n",
     "import matplotlib.pyplot as plt\n",
     "import numpy as np\n",
     "from IPython.display import HTML\n",
@@ -787,10 +784,8 @@
     "import torch.nn.functional as F\n",
     "import torch.utils.data as Data\n",
     "import torchvision\n",
-    "from sklearn.metrics import r2_score\n",
     "from torch import nn, optim\n",
     "from torch.autograd import Variable\n",
-    "from torch_lr_finder import LRFinder\n",
     "\n",
     "np.random.seed(14)  # For reproducibility\n",
     "torch.manual_seed(14)  # For reproducibility"

notebooks/Neural_network_for_Lorenz96.ipynb

@@ -27,7 +27,8 @@
     "from sklearn.metrics import r2_score\n",
     "from torch import nn, optim\n",
     "from torch.autograd import Variable\n",
-    "from torch_lr_finder import LRFinder\n",
+    "\n",
+    "# from torch_lr_finder import LRFinder # you might need to install the torch-lr-finder package\n",
     "\n",
     "np.random.seed(14)  # For reproducibility\n",
     "torch.manual_seed(14);  # For reproducibility"
@@ -1144,7 +1145,9 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Finding an optimal learning rate"
+    "## Finding an optimal learning rate\n",
+    "\n",
+    "To use the `LRFinder` package, uncomment the import of LRFinder in the top cell and then uncomment the next cell."
    ]
   },
   {
@@ -1153,12 +1156,12 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "nn_3l_LR = Net_ANN().double()\n",
-    "optimizer = optim.Adam(nn_3l_LR.parameters(), lr=1e-7)\n",
-    "lr_finder = LRFinder(nn_3l_LR, optimizer, criterion)\n",
-    "lr_finder.range_test(loader, end_lr=100, num_iter=200)\n",
-    "lr_finder.plot()  # to inspect the loss-learning rate graph\n",
-    "lr_finder.reset()  # to reset the model and optimizer to their initial state"
+    "# nn_3l_LR = Net_ANN().double()\n",
+    "# optimizer = optim.Adam(nn_3l_LR.parameters(), lr=1e-7)\n",
+    "# lr_finder = LRFinder(nn_3l_LR, optimizer, criterion)\n",
+    "# lr_finder.range_test(loader, end_lr=100, num_iter=200)\n",
+    "# lr_finder.plot()  # to inspect the loss-learning rate graph\n",
+    "# lr_finder.reset()  # to reset the model and optimizer to their initial state"
    ]
   },
   {
@@ -1167,19 +1170,19 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "n_epochs = 20  # Number of epocs\n",
-    "optimizer = optim.Adam(nn_3l_LR.parameters(), lr=0.01)\n",
-    "validation_loss = list()\n",
-    "train_loss = list()\n",
-    "# time0 = time()\n",
-    "for epoch in range(1, n_epochs + 1):\n",
-    "    train_model(nn_3l_LR, criterion, loader, optimizer)\n",
-    "    train_loss.append(test_model(nn_3l_LR, criterion, loader, optimizer, \"train\"))\n",
-    "    validation_loss.append(test_model(nn_3l_LR, criterion, loader_test, optimizer))\n",
-    "plt.plot(train_loss, \"b\", label=\"training loss\")\n",
-    "plt.plot(validation_loss, \"r\", label=\"validation loss\")\n",
+    "# n_epochs = 20  # Number of epocs\n",
+    "# optimizer = optim.Adam(nn_3l_LR.parameters(), lr=0.01)\n",
+    "# validation_loss = list()\n",
+    "# train_loss = list()\n",
+    "# # time0 = time()\n",
+    "# for epoch in range(1, n_epochs + 1):\n",
+    "#     train_model(nn_3l_LR, criterion, loader, optimizer)\n",
+    "#     train_loss.append(test_model(nn_3l_LR, criterion, loader, optimizer, \"train\"))\n",
+    "#     validation_loss.append(test_model(nn_3l_LR, criterion, loader_test, optimizer))\n",
+    "# plt.plot(train_loss, \"b\", label=\"training loss\")\n",
+    "# plt.plot(validation_loss, \"r\", label=\"validation loss\")\n",
     "\n",
-    "plt.legend();"
+    "# plt.legend();"
    ]
   },
   {
@@ -1195,11 +1198,11 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "optimizer = optim.Adam(nn_3l_LR.parameters(), lr=1e-7)\n",
-    "lr_finder = LRFinder(nn_3l_LR, optimizer, criterion)\n",
-    "lr_finder.range_test(loader, end_lr=100, num_iter=200)\n",
-    "lr_finder.plot()  # to inspect the loss-learning rate graph\n",
-    "lr_finder.reset()  # to reset the model and optimizer to their initial state"
+    "# optimizer = optim.Adam(nn_3l_LR.parameters(), lr=1e-7)\n",
+    "# lr_finder = LRFinder(nn_3l_LR, optimizer, criterion)\n",
+    "# lr_finder.range_test(loader, end_lr=100, num_iter=200)\n",
+    "# lr_finder.plot()  # to inspect the loss-learning rate graph\n",
+    "# lr_finder.reset()  # to reset the model and optimizer to their initial state"
    ]
   },
   {
@@ -1208,19 +1211,19 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "n_epochs = 10  # Number of epocs\n",
-    "optimizer = optim.Adam(nn_3l_LR.parameters(), lr=0.001)\n",
-    "validation_loss = list()\n",
-    "train_loss = list()\n",
-    "# time0 = time()\n",
-    "for epoch in range(1, n_epochs + 1):\n",
-    "    train_model(nn_3l_LR, criterion, loader, optimizer)\n",
-    "    train_loss.append(test_model(nn_3l_LR, criterion, loader, optimizer, \"train\"))\n",
-    "    validation_loss.append(test_model(nn_3l_LR, criterion, loader_test, optimizer))\n",
-    "plt.plot(train_loss, \"b\", label=\"training loss\")\n",
-    "plt.plot(validation_loss, \"r\", label=\"validation loss\")\n",
-    "\n",
-    "plt.legend();"
+    "# n_epochs = 10  # Number of epocs\n",
+    "# optimizer = optim.Adam(nn_3l_LR.parameters(), lr=0.001)\n",
+    "# validation_loss = list()\n",
+    "# train_loss = list()\n",
+    "# # time0 = time()\n",
+    "# for epoch in range(1, n_epochs + 1):\n",
+    "#     train_model(nn_3l_LR, criterion, loader, optimizer)\n",
+    "#     train_loss.append(test_model(nn_3l_LR, criterion, loader, optimizer, \"train\"))\n",
+    "#     validation_loss.append(test_model(nn_3l_LR, criterion, loader_test, optimizer))\n",
+    "# plt.plot(train_loss, \"b\", label=\"training loss\")\n",
+    "# plt.plot(validation_loss, \"r\", label=\"validation loss\")\n",
+    "\n",
+    "# plt.legend();"
    ]
   },
   {

notebooks/gradient_decent.ipynb

@@ -91,9 +91,7 @@
     "\n",
     "# from fastai.basics import *\n",
     "from matplotlib.animation import FuncAnimation\n",
-    "from torch import nn, optim\n",
-    "from torch.autograd import Variable\n",
-    "from torch_lr_finder import LRFinder\n",
+    "from torch import nn\n",
     "\n",
     "np.random.seed(42)"
    ]