Skip to content

Open-source speech recognition toolkit for training, inference, streaming ASR, VAD, punctuation, speaker diarization pipelines, and OpenAI-compatible/MCP serving.

update

游雁fffb628d31f8
diff --git a/funasr/train_utils/trainer.py b/funasr/train_utils/trainer.py
index 3e8358157..77eee60a5 100644
--- a/funasr/train_utils/trainer.py
+++ b/funasr/train_utils/trainer.py
@@ -248,8 +248,12 @@ class Trainer:
optim.zero_grad()
speed_stats = {}
time5 = time.perf_counter()
-
+ iterator_stop = torch.tensor(0).to(self.device)
for batch_idx, batch in enumerate(dataloader_train):
+ if self.use_ddp or self.use_fsdp:
+ dist.all_reduce(iterator_stop, dist.ReduceOp.SUM)
+ if iterator_stop > 0:
+ break
self.batch_total += 1
time1 = time.perf_counter()
speed_stats["data_load"] = f"{time1-time5:0.3f}"
@@ -356,7 +360,11 @@ class Trainer:
if (batch_idx+1) % self.save_checkpoint_interval == 0:
self.save_checkpoint(epoch, model=model, optim=optim, scheduler=scheduler, scaler=scaler, step=batch_idx+1)
-
+ else:
+ if self.use_ddp or self.use_fsdp:
+ iterator_stop.fill_(1)
+ dist.all_reduce(iterator_stop, dist.ReduceOp.SUM)
+
if self.use_ddp or self.use_fsdp:
dist.barrier()
@@ -383,7 +391,12 @@ class Trainer:
speed_stats = {}
time5 = time.perf_counter()
+ iterator_stop = torch.tensor(0).to(self.device)
for batch_idx, batch in enumerate(dataloader_val):
+ if self.use_ddp or self.use_fsdp:
+ dist.all_reduce(iterator_stop, dist.ReduceOp.SUM)
+ if iterator_stop > 0:
+ break
time1 = time.perf_counter()
speed_stats["data_load"] = f"{time1 - time5:0.3f}"
batch = to_device(batch, self.device)
@@ -433,9 +446,16 @@ class Trainer:
tag="val",
)
+ else:
+ if self.use_ddp or self.use_fsdp:
+ iterator_stop.fill_(1)
+ dist.all_reduce(iterator_stop, dist.ReduceOp.SUM)
+
self.val_acc_list.append(self.val_acc_avg)
model.train()
-
+
+
+
if self.use_ddp or self.use_fsdp:
dist.barrier()