From 4cf44a89f808411a0616c8ed92c3afae3d3e371a Mon Sep 17 00:00:00 2001
From: 游雁 <zhifu.gzf@alibaba-inc.com>
Date: 星期三, 21 二月 2024 00:43:27 +0800
Subject: [PATCH] bugfix
---
funasr/train_utils/trainer.py | 14 ++++++++------
1 files changed, 8 insertions(+), 6 deletions(-)
diff --git a/funasr/train_utils/trainer.py b/funasr/train_utils/trainer.py
index 10f7f80..b3c9953 100644
--- a/funasr/train_utils/trainer.py
+++ b/funasr/train_utils/trainer.py
@@ -188,7 +188,7 @@
epoch (int): The current epoch number.
"""
self.model.train()
- pbar = tqdm(colour="blue", desc=f"Training Epoch: {epoch + 1}", total=len(self.dataloader_train),
+ pbar = tqdm(colour="blue", desc=f"rank: {self.local_rank}, Training Epoch: {epoch + 1}", total=len(self.dataloader_train),
dynamic_ncols=True)
# Set the number of steps for gradient accumulation
@@ -273,12 +273,14 @@
torch.cuda.memory_reserved()/1024/1024/1024,
torch.cuda.max_memory_reserved()/1024/1024/1024,
)
+ lr = self.scheduler.get_last_lr()[0]
description = (
f"rank: {self.local_rank}, "
f"epoch: {epoch}/{self.max_epoch}, "
- f"step: {batch_idx}/{len(self.dataloader_train)}, total: {self.batch_total}, "
+ f"step: {batch_idx+1}/{len(self.dataloader_train)}, total: {self.batch_total}, "
f"(loss: {loss.detach().cpu().item():.3f}), "
- f"{[(k, round(v.cpu().item(), 3)) for k, v in stats.items()]}"
+ f"(lr: {lr:.3e}), "
+ f"{[(k, round(v.cpu().item(), 3)) for k, v in stats.items()]}, "
f"{speed_stats}, "
f"{gpu_info}"
)
@@ -307,7 +309,7 @@
"""
self.model.eval()
with torch.no_grad():
- pbar = tqdm(colour="red", desc=f"Training Epoch: {epoch + 1}", total=len(self.dataloader_val),
+ pbar = tqdm(colour="red", desc=f"rank: {self.local_rank}, Validation Epoch: {epoch + 1}", total=len(self.dataloader_val),
dynamic_ncols=True)
speed_stats = {}
time5 = time.perf_counter()
@@ -341,9 +343,9 @@
description = (
f"rank: {self.local_rank}, "
f"validation epoch: {epoch}/{self.max_epoch}, "
- f"step: {batch_idx}/{len(self.dataloader_val)}, "
+ f"step: {batch_idx+1}/{len(self.dataloader_val)}, "
f"(loss: {loss.detach().cpu().item():.3f}), "
- f"{[(k, round(v.cpu().item(), 3)) for k, v in stats.items()]}"
+ f"{[(k, round(v.cpu().item(), 3)) for k, v in stats.items()]}, "
f"{speed_stats}, "
)
pbar.set_description(description)
--
Gitblit v1.9.1