You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In the paper Neural Architectures for Named Entity Recognition, the author mentioned to use token at the beginning of a sequence, and a token at the end of a sequence. In your code I didn't find these two's info. According to a pytorch NER implementation,there are <start> and <end> tokens. By the way, in his implementation of viterbi decode,he take care of first step,to make sure it's <start>, but at tensorflow's code
trellis[0] = score[0]
for t in range(1, score.shape[0]):
v = np.expand_dims(trellis[t - 1], 1) + transition_params
trellis[t] = score[t] + np.max(v, 0)
backpointers[t] = np.argmax(v, 0)
I am not sure whether start end is necessary, and whether these two code are right?
The text was updated successfully, but these errors were encountered:
In the paper Neural Architectures for Named Entity Recognition, the author mentioned to use token at the beginning of a sequence, and a token at the end of a sequence. In your code I didn't find these two's info. According to a pytorch NER implementation,there are <start> and <end> tokens. By the way, in his implementation of
viterbi decode
,he take care of first step,to make sure it's <start>, but at tensorflow's codeI am not sure whether
start end
is necessary, and whether these two code are right?The text was updated successfully, but these errors were encountered: