ValueError: high is out of bounds for int32

experiencor / keras-yolo2

Easy training on custom dataset. Various backends (MobileNet and SqueezeNet) supported. A YOLO demo to detect raccoon run entirely in brower is accessible at https://git.io/vF7vI (not on Windows).

MIT License

1.73k stars 784 forks source link

ValueError: high is out of bounds for int32 #399

Open Twikkxx opened 5 years ago

Twikkxx commented 5 years ago

~\Anaconda3\lib\site-packages\imgaug-0.2.7-py3.6.egg\imgaug\imgaug.py in derive_random_states(randomstate, n) 430 431 """ --> 432 seed = random_state.randint(SEED_MIN_VALUE, SEED_MAX_VALUE, 1)[0] 433 return [new_randomstate(seed+i) for i in sm.xrange(n)] 434

mtrand.pyx in mtrand.RandomState.randint()

ValueError: high is out of bounds for int32

hi guys got this kind of error in the training part of the Yolo Step-by-Step (Setup a few callbacks and start the training)

mperini commented 5 years ago

Hi Twikkxx, not sure if that helps, I had a similar issue while running keras with imgaug, and it got solved by updating imgaug to the last version. I think there was some bug in the imgaug.py file, where they had defined (in the version I had downloaded) SEED_MAX_VALUE to be 232-1, while now it is 231-1.