Tutorials/AI-Workflows-Weights-and-Biases.md

@@ -65,5 +65,33 @@ for epoch in range(num_epochs):
 ```
 
 ## PyTorch Lightning Integration Example
+**Step 1: Import W&B logger**
 
-## Uploading Model Checkpoints to W&B
+```python
\ No newline at end of file
+from lightning.pytorch.loggers import WandbLogger
+```
+
+**Step 2: Initialize W&B logger**
+
+- Instantiate the `WandbLogger` and pass it to Lightning's `Trainer`. Make sure to save data locally during training by setting the mode to `offline`
+
+```python
+wandb_logger = WandbLogger(project="my_project", entity="my_entity", mode="offline")
+trainer = lightning.Trainer(**config["trainer"],
+                            logger=wandb_logger)
+```
+
+**Step 3: Training Step**
+
+```python
+def training_step(self, batch, batch_idx):
+    data, target = batch
+    output = self.forward(data)
+    loss = criterion(output, target)
+    # Log training metrics to W&B
+    self.log("train/loss", loss, on_step=False, on_epoch=True)
+    return loss       
+```
+
+## Model Checkpointing using W&B
+[TBD]
\ No newline at end of file