Bo Wang
Bo Wang
> This is where the concurrent eval job is useful: it saves the 3 best checkpoints (according to dev set) at any time. So in theory it should never save...
Thanks for your update! The strange thing is, it does work @abisee. By adding epsilon I have not encountered NaNs again. But it does affect the training convergence a bit....
@shahbazsyed Well I used as big a learning rate as 0.15, and had no NaN error after adding epsilon. How did you add the epsilon? Like this (as @rahul-iisc has...
@shahbazsyed I see you have changed the default parameter setting. I used 128 for `emb_dim`. And why did you increase your learning rate to 0.2 when you had NaN error...
Hi there, sorry for the slow reply. First of all you need to parse the data, I'd recommend using Stanford parser. Then you may use the `streamtwElec` function in https://github.com/bluemonk482/tdparse/blob/master/data/dataprocessing.py...
Hi @miloKnell The version CoNLL is same as the parsed data in https://github.com/bluemonk482/tdparse/tree/master/data/lidong/parses. I think I used the old Stanford parser, which gives more parsed type information than the CMU...
> ls -lh > total 148K > drwxrwxr-x 5 bowang bowang 4.0K Nov 20 18:10 configs > drwxrwxr-x 3 bowang bowang 4.0K Nov 20 18:21 data > -rw-rw-r-- 1 bowang...
@tmthyjames then you should edit the sampler.py file to make sure it is compatible with python3, e.g. `print('HierarchicalLDA sampling')`
Is there any plan on enabling multiprocessing for medspacy? Thanks!
That'd be great! Thanks @MichalMalyska !