PVNet 5.3.14__tar.gz → 5.3.16__tar.gz

{pvnet-5.3.14 → pvnet-5.3.16}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: PVNet
-Version: 5.3.14
+Version: 5.3.16
 Summary: PVNet
 Author-email: Peter Dudfield <info@openclimatefix.org>
 Requires-Python: <3.14,>=3.11

{pvnet-5.3.14 → pvnet-5.3.16}/PVNet.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: PVNet
-Version: 5.3.14
+Version: 5.3.16
 Summary: PVNet
 Author-email: Peter Dudfield <info@openclimatefix.org>
 Requires-Python: <3.14,>=3.11

{pvnet-5.3.14 → pvnet-5.3.16}/pvnet/training/lightning_module.py

@@ -206,23 +206,25 @@ class PVNetLightningModule(pl.LightningModule):
             if self.trainer.sanity_checking and plot_num == 0:
                 validate_batch_against_config(batch=batch, model=self.model)
 
-            with torch.no_grad():
+            # Save example forecast plots via logger
+            if self.logger:
                 y_hat = self.model(batch)
 
-            fig = plot_sample_forecasts(
-                batch,
-                y_hat,
-                quantiles=self.model.output_quantiles,
-                key_to_plot="generation",
-            )
+                fig = plot_sample_forecasts(
+                    batch,
+                    y_hat,
+                    quantiles=self.model.output_quantiles,
+                    key_to_plot="generation",
+                )
 
-            plot_name = f"val_forecast_samples/sample_set_{plot_num}"
+                plot_name = f"val_forecast_samples/sample_set_{plot_num}"
 
-            # Disabled for testing or using no logger
-            if self.logger:
-                self.logger.experiment.log({plot_name: wandb.Image(fig)})
+                self.logger.experiment.log(
+                    {plot_name: wandb.Image(fig)},
+                    step=self.trainer.global_step,
+                )
 
-            plt.close(fig)
+                plt.close(fig)
 
     def validation_step(self, batch: TensorBatch, batch_idx: int) -> None:
         """Run validation step"""
@@ -346,7 +348,10 @@ class PVNetLightningModule(pl.LightningModule):
                 title="Val horizon loss curve",
             )
 
-            wandb.log({"val_horizon_mae_plot": horizon_mae_plot})
+            wandb.log(
+                {"val_horizon_mae_plot": horizon_mae_plot},
+                step=self.trainer.global_step,                
+            )
 
             # Create persistence horizon accuracy curve but only on first epoch
             if self.current_epoch == 0:
@@ -357,4 +362,7 @@ class PVNetLightningModule(pl.LightningModule):
                     ylabel="MAE",
                     title="Val persistence horizon loss curve",
                 )
-                wandb.log({"persistence_val_horizon_mae_plot": persist_horizon_mae_plot})
+                wandb.log(
+                    {"persistence_val_horizon_mae_plot": persist_horizon_mae_plot},
+                    step=self.trainer.global_step,
+                )

{pvnet-5.3.14 → pvnet-5.3.16}/pvnet/training/plots.py

@@ -31,7 +31,8 @@ def plot_sample_forecasts(
     """Plot a batch of data and the forecast from that batch"""
 
     y = batch[key_to_plot].cpu().numpy()
-    y_hat = y_hat.cpu().numpy()    
+    y_hat = y_hat.cpu().numpy()
+    forecast_length = y_hat.shape[1]
     ids = batch["location_id"].cpu().numpy().squeeze()
     times_utc = pd.to_datetime(
         batch["time_utc"].cpu().numpy().squeeze().astype("datetime64[ns]")
@@ -42,30 +43,34 @@ def plot_sample_forecasts(
 
     for i, ax in enumerate(axes.ravel()[:batch_size]):
 
-        ax.plot(times_utc[i], y[i], marker=".", color="k", label=r"$y$")
+        # Crop to the forecast window only
+        forecast_times = times_utc[i][-forecast_length:]
+        y_no_history = y[i][-forecast_length:]
+
+        ax.plot(forecast_times, y_no_history, marker=".", color="k", label=r"$y$")
 
         if quantiles is None:
             ax.plot(
-                times_utc[i][-len(y_hat[i]) :],
-                y_hat[i], 
-                marker=".", 
-                color="r", 
+                forecast_times,
+                y_hat[i],
+                marker=".",
+                color="r",
                 label=r"$\hat{y}$",
             )
         else:
             cm = pylab.get_cmap("twilight")
             for nq, q in enumerate(quantiles):
                 ax.plot(
-                    times_utc[i][-len(y_hat[i]) :],
+                    forecast_times,
                     y_hat[i, :, nq],
                     color=cm(q),
                     label=r"$\hat{y}$" + f"({q})",
                     alpha=0.7,
                 )
 
-        ax.set_title(f"ID: {ids[i]} | {times_utc[i][0].date()}", fontsize="small")
+        ax.set_title(f"ID: {ids[i]} | {forecast_times[0].date()}", fontsize="small")
 
-        xticks = [t for t in times_utc[i] if t.minute == 0][::2]
+        xticks = [t for t in forecast_times if t.minute == 0][::2]
         ax.set_xticks(ticks=xticks, labels=[f"{t.hour:02}" for t in xticks], rotation=90)
         ax.grid()