kazutoshishinoda / nanhm-for-gec Goto Github PK

I try to use NUCLE to test this GEC's performance. When I run:

python Hybrid_BiGRU.py ./data/NUCLE/NUCLE_X_test.txt ./data/NUCLE/NUCLE_y_test.txt ./data/vocab/eng_source_word_vocab.txt ./data/vocab/eng_target_word_vocab.txt ./data/vocab/eng_source_char_vocab.txt ./data/vocab/eng_target_char_vocab.txt

I got this problems:

Traceback (most recent call last):
File "Hybrid_BiGRU.py", line 639, in
main()
File "Hybrid_BiGRU.py", line 618, in main
trainer.run()
File "/Users/codetracer/anaconda2/lib/python2.7/site-packages/chainer/training/trainer.py", line 313, in run
six.reraise(*sys.exc_info())
File "/Users/codetracer/anaconda2/lib/python2.7/site-packages/chainer/training/trainer.py", line 299, in run
update()
File "/Users/codetracer/anaconda2/lib/python2.7/site-packages/chainer/training/updater.py", line 223, in update
self.update_core()
File "/Users/codetracer/anaconda2/lib/python2.7/site-packages/chainer/training/updater.py", line 236, in update_core
optimizer.update(loss_func, **in_arrays)
File "/Users/codetracer/anaconda2/lib/python2.7/site-packages/chainer/optimizer.py", line 528, in update
loss = lossfun(*args, **kwds)
File "Hybrid_BiGRU.py", line 97, in call
loss, n_w, n_c, n_c_a = self.CalcLoss(xs, ys)
File "Hybrid_BiGRU.py", line 144, in CalcLoss
h_list, h_bar_list, c_s_list, z_s_list = self.decoder(None, ht, eys)
File "/Users/codetracer/NANHM-for-GEC/my_chainer/links/connection/n_step_gru.py", line 163, in call
self.n_layers, self.dropout, h0, hx, ws, bs, trans_x, W, B)
File "/Users/codetracer/NANHM-for-GEC/my_chainer/functions/connection/n_step_gru.py", line 124, in n_step_gru
use_bi_direction=False, **kwargs)
File "/Users/codetracer/NANHM-for-GEC/my_chainer/functions/connection/n_step_gru.py", line 466, in n_step_gru_base
hs, h_bars, c_ss, z_ss = _one_directional_loop(di=0)
File "/Users/codetrace/NANHM-for-GEC/my_chainer/functions/connection/n_step_gru.py", line 425, in _one_directional_loop
phi_ht, phi_d)) #(4)
File "/Users/codetracer/NANHM-for-GEC/my_chainer/functions/connection/n_step_gru.py", line 424, in
reshape.reshape((linear.linear(x, reshape.reshape(y,(1,len(y))))), (len(x),)),
TypeError: object of type 'NoneType' has no len()

I didn't figure out what was the problem here. Could you help to solve this problem? Thank you so much.

Hello, thanks for the replication of "A Nested Attention Neural Hybrid Model for Grammatical Error
Correction".

I am wondering that if the current codebase is still being maintained or is working or not.
Or could the code here replicates the result of the original paper?

Thanks a lot!