Improved model+EMA checkpointing 2 (#2295)

test.py

@@ -269,6 +269,7 @@ def test(data,
             print(f'pycocotools unable to run: {e}')
 
     # Return results
+    model.float()  # for training
     if not training:
         s = f"\n{len(list(save_dir.glob('labels/*.txt')))} labels saved to {save_dir / 'labels'}" if save_txt else ''
         print(f"Results saved to {save_dir}{s}")

train.py

@@ -4,6 +4,7 @@ import math
 import os
 import random
 import time
+from copy import deepcopy
 from pathlib import Path
 from threading import Thread
 
@@ -381,8 +382,8 @@ def train(hyp, opt, device, tb_writer=None, wandb=None):
                 ckpt = {'epoch': epoch,
                         'best_fitness': best_fitness,
                         'training_results': results_file.read_text(),
-                        'model': (model.module if is_parallel(model) else model).half(),
-                        'ema': (ema.ema.half(), ema.updates),
+                        'model': deepcopy(model.module if is_parallel(model) else model).half(),
+                        'ema': (deepcopy(ema.ema).half(), ema.updates),
                         'optimizer': optimizer.state_dict(),
                         'wandb_id': wandb_run.id if wandb else None}
 
@@ -392,8 +393,6 @@ def train(hyp, opt, device, tb_writer=None, wandb=None):
                     torch.save(ckpt, best)
                 del ckpt
 
-            model.float(), ema.ema.float()
-
         # end epoch ----------------------------------------------------------------------------------------------------
     # end training