Closed yongzhuo closed 5 years ago
@yongzhuo 貌似是cuda的问题,试试cpu版本
where i change tf to keras. it is ok ! project: https://github.com/CyberZHG/keras-bert demo: https://github.com/CyberZHG/keras-bert/blob/master/demo/load_model/load_and_pool.py I think there was something with the graph of queue, maybe it is wrong with shareing the graph. or we couldn't use queue it in one single gpu? when using keras called by django, first it wasn't work for "tensor *** is not an element in the graph." we use "global model"、"global gragh" , it success!
@yongzhuo 卓哥,你真调皮
Floating point exception (core dumped) when running extract_feature.py using linux、tensorflow-gpu==1.12.0 and SystemError: error return without exception set when debug with windows 10. zhao, what we need is your help。。。。。。 crying