open-sesame
open-sesame copied to clipboard
swabhs
A frame-semantic parsing system based on a softmax-margin SegRNN.
Hi, Thanks for this work. I was trying to test on unannoatated data using the predict model. But, when I try to run the targetid, it throws following error. I...
The pre-trained models are incompatible with official GloVe mebddings. Re-training frameid ends up with -nan gradients at the early stage of training.
Hello. Thanks for all your work. When I try to train the model for FrameID and ArgID, I get the following key error. Would you be able to help me?...
Is the reported frameid result (89.74 on dev and 86.55 on test) using the gold target labels? And what's the frameid result if targetid module is used instead of gold...
C:\Users\Saravanan>python -m sesame.preprocess Traceback (most recent call last): File "C:\Users\Saravanan\AppData\Local\Programs\Python\Python38\lib\runpy.py", line 192, in _run_module_as_main return _run_code(code, main_globals, None, File "C:\Users\Saravanan\AppData\Local\Programs\Python\Python38\lib\runpy.py", line 85, in _run_code exec(code, run_globals) File "C:\Users\Saravanan\AppData\Local\Programs\Python\Python38\sesame\preprocess.py", line 31,...
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,...
I run `python -m sesame.targetid --mode predict --model_name fn1.7-pretrained-targetid --raw_input sentences.txt` and find the ouput in `logs/fn1.7-pretrained-targetid/predicted-targets.conll` is total blank. It turned out that `instance.postags` and `instance.lemmas` are all lists...
