W&B ID reset on training completion (ultralytics#1852)

* Update train.py Fix the bug of always the same W&B ID and continue overwrite with the old logging. BUG report ultralytics#1851 * Fix the bug of duplicate W&B ID fix the bug of ultralytics#1851 If we had trained on yolov5s.pt, the program will generate a new unique W&B ID. If we hadn't, the program will keep the old code, we can still use --resume aug. * Update general.py * revert train.py changes Co-authored-by: Glenn Jocher <[email protected]>
taicaile · Jan 7, 2021 · 405f3a3 · 405f3a3
1 parent a92320c
commit 405f3a3
Showing 1 changed file with 2 additions and 2 deletions.
diff --git a/utils/general.py b/utils/general.py
@@ -361,8 +361,8 @@ def non_max_suppression(prediction, conf_thres=0.25, iou_thres=0.45, classes=Non
 def strip_optimizer(f='weights/best.pt', s=''):  # from utils.general import *; strip_optimizer()
     # Strip optimizer from 'f' to finalize training, optionally save as 's'
     x = torch.load(f, map_location=torch.device('cpu'))
-    x['optimizer'] = None
-    x['training_results'] = None
+    for key in 'optimizer', 'training_results', 'wandb_id':
+        x[key] = None
     x['epoch'] = -1
     x['model'].half()  # to FP16
     for p in x['model'].parameters():