xclele/yolov10
1
0
Fork 0
mirror of https://github.com/THU-MIG/yolov10.git synced 2025-05-24 06:14:55 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

fix for KeyError: 'epoch'

Browse Source
This commit is contained in:
Thomas Friedel 2024-05-26 02:03:08 +02:00
parent f19f3e521f
commit e369c45a62
1 changed files with 4 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
ultralytics/engine/trainer.py
View File

@ -224,10 +224,9 @@ class BaseTrainer:
"""Initializes and sets the DistributedDataParallel parameters for training.""" """Initializes and sets the DistributedDataParallel parameters for training."""
torch.cuda.set_device(RANK) torch.cuda.set_device(RANK)
self.device = torch.device("cuda", RANK) self.device = torch.device("cuda", RANK)
# LOGGER.info(f'DDP info: RANK {RANK}, WORLD_SIZE {world_size}, DEVICE {self.device}') os.environ["TORCH_NCCL_BLOCKING_WAIT"] = "1" # set to enforce timeout
os.environ["NCCL_BLOCKING_WAIT"] = "1" # set to enforce timeout
dist.init_process_group( dist.init_process_group(
"nccl" if dist.is_nccl_available() else "gloo", backend="nccl" if dist.is_nccl_available() else "gloo",
timeout=timedelta(seconds=10800), # 3 hours timeout=timedelta(seconds=10800), # 3 hours
rank=RANK, rank=RANK,
world_size=world_size, world_size=world_size,
@ -659,7 +658,8 @@ class BaseTrainer:
def resume_training(self, ckpt): def resume_training(self, ckpt):
"""Resume YOLO training from given epoch and best fitness.""" """Resume YOLO training from given epoch and best fitness."""
if ckpt is None: print(f"resume_training: {ckpt}")
if ckpt is None or not self.resume:
return return
best_fitness = 0.0 best_fitness = 0.0
start_epoch = ckpt["epoch"] + 1 start_epoch = ckpt["epoch"] + 1