done with the Python part of model serialization tutorial

Author: motus

Diff for: load/state_dict.ipynb

@@ -140,7 +140,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "Remember, that initially all parameters are initialized with random values, e.g."
+    "Remember, that after the model instantiation its parameters are initialized with random values, e.g."
    ]
   },
   {
@@ -152,8 +152,8 @@
      "data": {
       "text/plain": [
        "Parameter containing:\n",
-       "tensor([ 0.1578,  0.1650, -0.1272, -0.1976,  0.0318, -0.1246, -0.0474, -0.0620,\n",
-       "         0.1829, -0.1198], requires_grad=True)"
+       "tensor([-0.1677,  0.1930, -0.0675,  0.0393, -0.1062, -0.0804,  0.0219,  0.0191,\n",
+       "         0.0188, -0.1543], requires_grad=True)"
       ]
      },
      "execution_count": 4,
@@ -175,7 +175,9 @@
   {
    "cell_type": "code",
    "execution_count": 5,
-   "metadata": {},
+   "metadata": {
+    "scrolled": true
+   },
    "outputs": [
     {
      "data": {
@@ -196,6 +198,15 @@
     "model.conv1.bias"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Now we can use our model for scoring.\n",
+    "\n",
+    "To continue training from the saved checkpoint, we also need to instantiate the optimizer and initialize it (using [`torch.optim.Optimizer.load_state_dict()`](https://pytorch.org/docs/stable/optim.html#torch.optim.Optimizer.load_state_dict)). This is required because many of the optimizer's parameters (learning rate, momentum, etc.) change in training, and we often want to resume training from a certain checkpoint."
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,