W&B: track batch size after autobatch

train.py

@@ -138,6 +138,7 @@ def train(hyp,  # path/to/hyp.yaml or hyp dictionary
     # Batch size
     if RANK == -1 and batch_size == -1:  # single-GPU only, estimate best batch size
         batch_size = check_train_batch_size(model, imgsz)
+        loggers.on_params_update({"batch_size": batch_size})
 
     # Optimizer
     nbs = 64  # nominal batch size

utils/callbacks.py

@@ -32,7 +32,7 @@ def __init__(self):
             'on_fit_epoch_end': [],  # fit = train + val
             'on_model_save': [],
             'on_train_end': [],
-
+            'on_params_update': [],
             'teardown': [],
         }
 

utils/loggers/__init__.py

@@ -157,3 +157,9 @@ def on_train_end(self, last, best, plots, epoch, results):
             else:
                 self.wandb.finish_run()
                 self.wandb = WandbLogger(self.opt)
+
+    def on_params_update(self, params):
+        # Update hyperparams or configs of the experiment
+        # params: A dict containing {param: value} pairs
+        if self.wandb:
+            self.wandb.wandb_run.config.update(params, allow_val_change=True)