From 26ab38b56cb4f69cc82e5d58907f2a57f6f2cbdd Mon Sep 17 00:00:00 2001
From: 游雁 <zhifu.gzf@alibaba-inc.com>
Date: 星期三, 21 二月 2024 14:22:14 +0800
Subject: [PATCH] update train recipe
---
funasr/train_utils/trainer.py | 27 +++++++++++++++++++++------
1 files changed, 21 insertions(+), 6 deletions(-)
diff --git a/funasr/train_utils/trainer.py b/funasr/train_utils/trainer.py
index f375384..f4ef773 100644
--- a/funasr/train_utils/trainer.py
+++ b/funasr/train_utils/trainer.py
@@ -3,6 +3,7 @@
import torch
import logging
from tqdm import tqdm
+from datetime import datetime
import torch.distributed as dist
from contextlib import nullcontext
# from torch.utils.tensorboard import SummaryWriter
@@ -109,12 +110,8 @@
print(f'Checkpoint saved to {filename}')
latest = Path(os.path.join(self.output_dir, f'model.pt'))
- try:
- latest.unlink()
- except:
- pass
+ torch.save(state, latest)
- latest.symlink_to(filename)
def _resume_checkpoint(self, resume_path):
"""
@@ -128,7 +125,20 @@
if os.path.isfile(ckpt):
checkpoint = torch.load(ckpt)
self.start_epoch = checkpoint['epoch'] + 1
- self.model.load_state_dict(checkpoint['state_dict'])
+ # self.model.load_state_dict(checkpoint['state_dict'])
+ src_state = checkpoint['state_dict']
+ dst_state = self.model.state_dict()
+ for k in dst_state.keys():
+ if not k.startswith("module.") and "module."+k in src_state.keys():
+ k_ddp = "module."+k
+ else:
+ k_ddp = k
+ if k_ddp in src_state.keys():
+ dst_state[k] = src_state[k_ddp]
+ else:
+ print(f"Miss key in ckpt: model: {k}, ckpt: {k_ddp}")
+
+ self.model.load_state_dict(dst_state)
self.optim.load_state_dict(checkpoint['optimizer'])
self.scheduler.load_state_dict(checkpoint['scheduler'])
print(f"Checkpoint loaded successfully from '{ckpt}'")
@@ -273,11 +283,16 @@
torch.cuda.memory_reserved()/1024/1024/1024,
torch.cuda.max_memory_reserved()/1024/1024/1024,
)
+ lr = self.scheduler.get_last_lr()[0]
+ time_now = datetime.now()
+ time_now = now.strftime("%Y-%m-%d %H:%M:%S")
description = (
+ f"{time_now}, "
f"rank: {self.local_rank}, "
f"epoch: {epoch}/{self.max_epoch}, "
f"step: {batch_idx+1}/{len(self.dataloader_train)}, total: {self.batch_total}, "
f"(loss: {loss.detach().cpu().item():.3f}), "
+ f"(lr: {lr:.3e}), "
f"{[(k, round(v.cpu().item(), 3)) for k, v in stats.items()]}, "
f"{speed_stats}, "
f"{gpu_info}"
--
Gitblit v1.9.1