问题遇到的现象和发生背景
我做实例分割训练自己的数据集,训练次数MAX_ITER: 200会出现错误,要把次数提到上千次才行,但是gpu这个时候就受不了了,这是什么原因?要怎么修改?
问题相关代码,请勿粘贴截图
运行结果及报错内容
Traceback (most recent call last):
File "tools/train_net.py", line 226, in <module>
launch(
File "f:\cocoapi\detectron2-0.4\detectron2\engine\launch.py", line 62, in launch
main_func(*args)
File "tools/train_net.py", line 214, in main
trainer = Trainer(cfg)
File "f:\cocoapi\detectron2-0.4\detectron2\engine\defaults.py", line 323, in __init__
self.scheduler = self.build_lr_scheduler(cfg, optimizer)
File "f:\cocoapi\detectron2-0.4\detectron2\engine\defaults.py", line 474, in build_lr_scheduler
return build_lr_scheduler(cfg, optimizer)
File "f:\cocoapi\detectron2-0.4\detectron2\solver\build.py", line 246, in build_lr_scheduler
sched = WarmupParamScheduler(
File "f:\cocoapi\detectron2-0.4\detectron2\solver\lr_scheduler.py", line 37, in __init__
end_value = scheduler(warmup_length) # the value to reach when warmup ends
File "E:\Anaconda\envs\detectron2\lib\site-packages\fvcore\common\param_scheduler.py", line 236, in __call__
raise RuntimeError(
RuntimeError: where in ParamScheduler must be in [0, 1]: got 5.0