concern on sentence

[niranjan8129]

I liked your model . i have concern .does it generate sentence ?

eg
input : fever
output : i am having fever since last night

[stardut]

This model forms a sentence by predicting the next word in the previous article. In other words, the input word will be the first word of the sentence, not in the sentence.

I think this sence is match your concern:

input: fever
output: fever i am having fever since last night.

the input word is the fisrt word in the generate sentence and the model is work.

[niranjan8129]

Thanks for your quick response . I will try now . one more help .. how to create vocab.pkl file ? could you please upload vocab.pkl model/source code if you have ?

[stardut]

The vocab.pkl was created by the training data file tang in /data/tang while running train.py. You need to modify the training data file to your train data in train.py and sample.py :

data_dir = 'data/'
input_file = 'tang' #change to your train data file path. Each row represents a training data in the training file.
vocab_file = 'vocab.pkl'  # create when run train.py
tensor_file = 'tensor.npy' # create when run train.py

[niranjan8129]

Thank you friend from bottom of my heart.. i am extremely delighted for your prompt response . Please clear me If i am not wrong. both tang and poetry data sets are different .. right?

Also, below text variables are input text.
text1 = '两个黄鹂鸣翠柳，'
text2 = '<《春》 '
text3 = '<'

pleases drop a text to niranjan8129@gmail.com if you dont feel bad.

Thanks
Niranjan

[stardut]

Yes, you are right.

[niranjan8129]

Hi friend , wish you a happy new year to you and your family . I was trying to train the model but it did not complete after running 26 hours with high end machine . Could you please help me the issue ?

[stardut]

Happy new year. The model net will be run forever and the model will be saved every 1000 step. So you can use it do something you want.

[niranjan8129]

Hi friend , I tried but could not figure out how to stop/completed after some batchsize/epoch . Could you please help me where I need to change the code if you dont mind

[frostbite-zhs]

Hi friend , I tried but could not figure out how to stop/completed after some batchsize/epoch . Could you please help me where I need to change the code if you dont mind

Hi friend , I tried but could not figure out how to stop/completed after some batchsize/epoch . Could you please help me where I need to change the code if you dont mind

你不需要把整个epoch跑完，只要有模型保存就可以使用了，但是如果没有训练完的话跑出来的效果预计不好。具体操作就是：运行train.py以后会生成一个model文件夹里面就保存了训练出来的模型，名字大致长这个样子model_loss_2.0342.ckpt-587000加上后缀名，之后就在sample.py文件大概33行指定模型，具体代码就是
#saver.restore(self.sess, 'model/model')
#saver.restore(self.sess, tf.train.latest_checkpoint("model/"))
saver.restore(self.sess, "model/model_loss_2.0342.ckpt-587000")
其中model_loss_2.0342.ckpt-587000是我跑出来的已经保存在model文件夹的模型
正如原作者所说，不需要跑完所有的epoch，只要有模型保存，就可以拿来使用了！