Moving images to the _static directory.

Author: drewoldag

docs/notebooks/results_dir.png docs/_static/screenshots/results_dir.pngdocs/notebooks/results_dir.png renamed to docs/_static/screenshots/results_dir.png

@@ -23,7 +23,7 @@
     "\n",
     "Whenever training or inference is completed, a `results` directory is generated that contains timestamped \"train\" and \"infer\" folders\n",
     "\n",
-    "![results_dir.png](results_dir.png)"
+    "![results_dir.png](../_static/screenshots/results_dir.png)"
    ]
   },
   {
@@ -126,7 +126,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.12.9"
+   "version": "3.13.12"
   }
  },
  "nbformat": 4,

…uted/tensorboard_checkpoint_val_loss.png …hots/tensorboard_checkpoint_val_loss.pngdocs/pre_executed/tensorboard_checkpoint_val_loss.png renamed to docs/_static/screenshots/tensorboard_checkpoint_val_loss.png docs/pre_executed/tensorboard_checkpoint_val_loss.png renamed to docs/_static/screenshots/tensorboard_checkpoint_val_loss.png

@@ -120,7 +120,7 @@
    "source": [
     "The resumed training data will create its own timestamped results directory with new checkpoints and will not overwrite the previous one. Note that if a model that should run for `n` epochs finishes its training and the checkpoint is loaded with the number of epochs still being `n`, then the model will be trained for `n` more epochs. This can be seen in TensorBoard as well; the validation loss is plotted against each epoch and the red curve represents the validation loss from the initial training while the blue curve is the validation loss from the additional 3 epochs.\n",
     "\n",
-    "![tensorboard_checkpoint_val_loss.png](tensorboard_checkpoint_val_loss.png)"
+    "![tensorboard_checkpoint_val_loss.png](../_static/screenshots/tensorboard_checkpoint_val_loss.png)"
    ]
   }
  ],
@@ -140,7 +140,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.12.12"
+   "version": "3.13.12"
   }
  },
  "nbformat": 4,

…ecuted/tensorboard_lr_history_cosine.png …nshots/tensorboard_lr_history_cosine.pngdocs/pre_executed/tensorboard_lr_history_cosine.png renamed to docs/_static/screenshots/tensorboard_lr_history_cosine.png docs/pre_executed/tensorboard_lr_history_cosine.png renamed to docs/_static/screenshots/tensorboard_lr_history_cosine.png

@@ -271,7 +271,7 @@
    "source": [
     "The results may be a bit worse than without using the learning rate scheduler, but that is to be expected since we didn't really try doing any sort of tuning and just wanted to demonstrate the feature. We can also see the history of the learning rates in TensorBoard to verify that the scheduler did in fact have an effect. Orange curves belong to the original model and red curves pertain to the model with the updated scheduler:\n",
     "\n",
-    "![tensorboard_lr_history_exponential.png](tensorboard_lr_history_exponential.png)"
+    "![tensorboard_lr_history_exponential.png](../_static/screenshots/tensorboard_lr_history_exponential.png)"
    ]
   },
   {
@@ -364,7 +364,7 @@
    "source": [
     "These results are also worse than before since we didn't do any tuning; the goal is simply to demonstrate using schedulers as a feature. The new pink curves on Tensorboard are from the model using CosineAnnealingLR:\n",
     "\n",
-    "![tensorboard_lr_history_cosine.png](tensorboard_lr_history_cosine.png)"
+    "![tensorboard_lr_history_cosine.png](../_static/screenshots/tensorboard_lr_history_cosine.png)"
    ]
   },
   {
@@ -394,7 +394,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.12.12"
+   "version": "3.13.12"
   }
  },
  "nbformat": 4,