-
https://github.com/tannerbohn/char2vec/blob/master/main.py#L101
Not at all obvious why should one subtract 0.5 from 1-hot vectors in X ?
-
Thank you very much for your work！
Can you provide the code for generating file **bert_base_cosine_adj.pkl_emb**, file **char2vec_cosine_adj.pkl_emb** and file **glove_cosine_adj.pkl_emb** ?
I th…
-
I referred to the website where I found a piece of code to reproduce training of Char2Vec but it produces an error, can you please help. If it works I want to train it for my purpose on the german tex…
-
Q1: What is windowed LSTM model?Is it the model that[ this paper ](http://iiis.tsinghua.edu.cn/~weblt/papers/window-lstm-morph-segmentation.pdf) presented?
Q2: How to transform the raw data li…
-
elif embedd_dim + 1 != len(tokens):
第一行是个数和维度，所以len(tokens)==2，但是后面都是301个元素，不太对吧
我修改以后，运行程序报了个：
ERROR: LOSS EXPLOSION (>1e8) ! PLEASE SET PROPER PARAMETERS AND STRUCTURE! EXIT....
也不知道到底是哪…
-
-
Hello,
I was trying to run the code:
=====================
dropout_keep_prob = tf.placeholder_with_default(1.0, (), name='keep_prob')
if baseline:
c2v = BasicEmbedding(model_params, vocab_siz…
-
I am using Tensorflow 1.4 and I get the error below:
33000 examples loaded
max word len 42
Traceback (most recent call last):
File "langid.py", line 121, in
dropout_keep_prob=dropout_kee…
-
Hello,
I'm trying to use your code with wikipedia corpus and a gensim model with vector size 400.
I've modified the code inside word_segmenter.py to have a layer size of 400 (replacing all 256)
b…