Update solver.py

Fixed the incorrect calculation of the epoch number when diffusion training was resumed from an existing model checkpoint.

diffusion/solver.py

@@ -101,6 +101,7 @@ def train(args, initial_global_step, model, optimizer, scheduler, vocoder, loade
     
     # run
     num_batches = len(loader_train)
+    start_epoch = initial_global_step // num_batches
     model.train()
     saver.log_info('======= start training =======')
     scaler = GradScaler()
@@ -113,7 +114,7 @@ def train(args, initial_global_step, model, optimizer, scheduler, vocoder, loade
     else:
         raise ValueError(' [x] Unknown amp_dtype: ' + args.train.amp_dtype)
     saver.log_info("epoch|batch_idx/num_batches|output_dir|batch/s|lr|time|step")
-    for epoch in range(args.train.epochs):
+    for epoch in range(start_epoch, args.train.epochs):
         for batch_idx, data in enumerate(loader_train):
             saver.global_step_increment()
             optimizer.zero_grad()