feature: fixed training loop arguments for Lightning module

notebooks/trainer.ipynb

@@ -4,7 +4,62 @@
    "cell_type": "code",
    "execution_count": null,
    "outputs": [],
-   "source": [],
+   "source": [
+    "import torch.optim\n",
+    "import pytorch_lightning as pl"
+   ],
+   "metadata": {
+    "collapsed": false,
+    "pycharm": {
+     "name": "#%%\n"
+    }
+   }
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "outputs": [],
+   "source": [
+    "class LitTrainer(pl.LightningModule):\n",
+    "    def __init__(self, model, loss_fn, optim):\n",
+    "        super().__init__()\n",
+    "        self.model = model\n",
+    "        self.loss_fn = loss_fn\n",
+    "        self.optim = optim\n",
+    "\n",
+    "    def training_step(self, batch, batch_idx):\n",
+    "        x, y = batch\n",
+    "        x = x.to(torch.float32)\n",
+    "\n",
+    "        y_pred = self.model(x).reshape(1, -1)\n",
+    "        train_loss = self.loss_fn(y_pred, y)\n",
+    "\n",
+    "        self.log(\"train_loss\", train_loss)\n",
+    "        return train_loss\n",
+    "\n",
+    "    def validation_step(self, batch, batch_idx):\n",
+    "        # this is the validation loop\n",
+    "        x, y = batch\n",
+    "        x = x.to(torch.float32)\n",
+    "\n",
+    "        y_pred = self.model(x).reshape(1, -1)\n",
+    "        validate_loss = self.loss_fn(y_pred, y)\n",
+    "\n",
+    "        self.log(\"val_loss\", validate_loss)\n",
+    "\n",
+    "    def test_step(self, batch, batch_idx):\n",
+    "        # this is the test loop\n",
+    "        x, y = batch\n",
+    "        x = x.to(torch.float32)\n",
+    "\n",
+    "        y_pred = self.model(x).reshape(1, -1)\n",
+    "        test_loss = self.loss_fn(y_pred, y)\n",
+    "\n",
+    "        self.log(\"test_loss\", test_loss)\n",
+    "\n",
+    "    def configure_optimizers(self):\n",
+    "        return self.optim\n"
+   ],
    "metadata": {
     "collapsed": false,
     "pycharm": {

src/trainer.py

@@ -1,22 +1,22 @@
 #!/usr/bin/env python
 # coding: utf-8
-import torch.optim
+import torch
+from torch import nn, optim
 import pytorch_lightning as pl
 
 
 class LitTrainer(pl.LightningModule):
-    def __init__(self, model, loss_fn, optim):
+    def __init__(self, model):
         super().__init__()
         self.model = model
-        self.loss_fn = loss_fn
-        self.optim = optim
+        self.optim = optim.Adam(self.parameters(), lr=1e-4)
+        self.loss = nn.CrossEntropyLoss()
 
     def training_step(self, batch, batch_idx):
         x, y = batch
-        x = x.to(torch.float32)
 
         y_pred = self.model(x).reshape(1, -1)
-        train_loss = self.loss_fn(y_pred, y)
+        train_loss = self.loss(y_pred, y)
 
         self.log("train_loss", train_loss)
         return train_loss
@@ -24,22 +24,23 @@ class LitTrainer(pl.LightningModule):
     def validation_step(self, batch, batch_idx):
         # this is the validation loop
         x, y = batch
-        x = x.to(torch.float32)
 
         y_pred = self.model(x).reshape(1, -1)
-        validate_loss = self.loss_fn(y_pred, y)
+        validate_loss = self.loss(y_pred, y)
 
         self.log("val_loss", validate_loss)
 
     def test_step(self, batch, batch_idx):
         # this is the test loop
         x, y = batch
-        x = x.to(torch.float32)
 
         y_pred = self.model(x).reshape(1, -1)
-        test_loss = self.loss_fn(y_pred, y)
+        test_loss = self.loss(y_pred, y)
 
         self.log("test_loss", test_loss)
 
+    def forward(self, x):
+        return self.model(x)
+
     def configure_optimizers(self):
         return self.optim