1) is there a resume function to training that enables one to not start from scratch each time?
2) approximately how long should I expect to train each network frame, target and arg on a GPU with 3GB memory?
3) can you point to any visualization tools for looking at the training data and predicted data? I think they are all in CONLL2009 format
4) can you recommend a good way think about the number of epochs per training round? How was 5 chosen?
5) why does predict mode for frameid.py require me to read the training data?
1) is there a resume function to training that enables one to not start from scratch each time? 2) approximately how long should I expect to train each network frame, target and arg on a GPU with 3GB memory? 3) can you point to any visualization tools for looking at the training data and predicted data? I think they are all in CONLL2009 format 4) can you recommend a good way think about the number of epochs per training round? How was 5 chosen? 5) why does predict mode for
frameid.pyrequire me to read the training data?Thanks --aaron