⬆️ [Update] mAP from 0~1 to percentage

yolo/tools/solver.py

@@ -247,7 +247,7 @@ class ModelValidator:
                     for mAP_key, mAP_val in mAP.items():
                         mAPs[mAP_key].append(mAP_val)
 
-            avg_mAPs = {key: torch.mean(torch.stack(val)) for key, val in mAPs.items()}
+            avg_mAPs = {key: 100 * torch.mean(torch.stack(val)) for key, val in mAPs.items()}
             self.progress.one_batch(avg_mAPs)
 
             predict_json.extend(predicts_to_json(img_paths, predicts, rev_tensor))

yolo/utils/logging_utils.py

@@ -117,6 +117,7 @@ class ProgressLogger(Progress):
     @rank_check
     def start_train(self, num_epochs: int):
         self.task_epoch = self.add_task(f"[cyan]Start Training {num_epochs} epochs", total=num_epochs)
+        self.update(self.task_epoch, advance=-0.5)
 
     @rank_check
     def start_one_epoch(
@@ -218,7 +219,7 @@ class ProgressLogger(Progress):
 
     @rank_check
     def finish_pycocotools(self, result, epoch_idx=-1):
-        ap_table, ap_main = make_ap_table(result, self.ap_past_list, self.last_result, epoch_idx)
+        ap_table, ap_main = make_ap_table(result * 100, self.ap_past_list, self.last_result, epoch_idx)
         self.last_result = np.maximum(result, self.last_result)
         self.ap_past_list.append((epoch_idx, ap_main))
         self.ap_table = ap_table

yolo/utils/solver_utils.py

@@ -21,9 +21,9 @@ def make_ap_table(score, past_result=[], last_score=None, epoch=-1):
     ap_table = Table()
     ap_table.add_column("Epoch", justify="center", style="white", width=5)
     ap_table.add_column("Avg. Precision", justify="left", style="cyan")
-    ap_table.add_column("", justify="right", style="green", width=5)
+    ap_table.add_column("%", justify="right", style="green", width=5)
     ap_table.add_column("Avg. Recall", justify="left", style="cyan")
-    ap_table.add_column("", justify="right", style="green", width=5)
+    ap_table.add_column("%", justify="right", style="green", width=5)
 
     for eps, (ap_name1, ap_color1, ap_value1, ap_name2, ap_color2, ap_value2) in past_result:
         ap_table.add_row(f"{eps: 3d}", ap_name1, f"{ap_color1}{ap_value1:.2f}", ap_name2, f"{ap_color2}{ap_value2:.2f}")