GPU训练yolov5问题

Traceback (most recent call last):
File "train.py", line 641, in <module>
main_worker(opt.device_id, ndevs_per_node, opt)
File "train.py", line 189, in main_worker
model = Model(opt.cfg)
File "/torch/src/pytorch_models/Training/Detection/yolov5s/models/yolo.py", line 70, in __init__
self._initialize_biases() # only run once
File "/torch/src/pytorch_models/Training/Detection/yolov5s/models/yolo.py", line 121, in _initialize_biases
b[:, 4] += math.log(8 / (640 / s) ** 2) # obj (8 objects per 640 image)
RuntimeError: a view of a leaf Variable that requires grad is being used in an in-place operation.

在报错的这一块增加 with torch.no_grad():