Change metrics to global loss (default) and also allow summary-net or…

… inference-net specific metrics for future

bayesflow/experimental/amortizers/amortizer.py

@@ -81,6 +81,36 @@ def reset_metrics(self):
     def metrics(self):
         return [self.loss_tracker]
 
+    def compute_metrics(self, x: dict, y: dict, y_pred: dict, **kwargs):
+        inferred_variables = self.configure_inferred_variables(x)
+        observed_variables = self.configure_observed_variables(x)
+
+        if self.summary_network:
+            summary_conditions = self.configure_summary_conditions(x)
+            summary_metrics = self.summary_network.compute_metrics(
+                x=(observed_variables, summary_conditions),
+                y=y.get("summary_targets"),
+                y_pred=y_pred.get("summary_outputs")
+            )
+        else:
+            summary_metrics = {}
+
+        inference_conditions = self.configure_inference_conditions(x, y_pred.get("summary_outputs"))
+        inference_metrics = self.inference_network.compute_metrics(
+            x=(inferred_variables, inference_conditions),
+            y=y.get("inference_targets"),
+            y_pred=y_pred.get("inference_outputs")
+        )
+
+        summary_metrics = {f"summary/{key}": value for key, value in summary_metrics.items()}
+        inference_metrics = {f"inference/{key}": value for key, value in inference_metrics.items()}
+
+        out_metrics = {'loss': self.loss_tracker.total.value}
+        out_metrics = out_metrics | summary_metrics
+        out_metrics = out_metrics | inference_metrics
+
+        return out_metrics
+
     def sample(self, data: dict, num_samples: int, sample_summaries=False, **kwargs):
 
         # Configure everything -> inference conditions / summary conditions