any guide to get a quick start ?

Question

any guide to get a quick start ?

SeekPoint opened this issue 7 years ago · comments

yuanke commented 7 years ago

plug_and_play.pickle

yuanke · Answer 1 · Tue Dec 19 2017 16:56:46 GMT+0800 (China Standard Time)

where is the plug_and_play.pickle

Animesh Karnewar · Answer 2 · Wed Dec 20 2017 00:34:04 GMT+0800 (China Standard Time)

I am working on the readme.

Animesh Karnewar · Answer 3 · Wed Dec 20 2017 00:35:45 GMT+0800 (China Standard Time)

To generate the plug_and_play.pickle, you need to run the fast_preprocessing_part1 and part2 scripts.
Also, download the wiko-biography dataset and put the train.nb, train.sent and train.box files in the data folder.

yuanke · Answer 4 · Mon Sep 03 2018 09:30:00 GMT+0800 (China Standard Time)

waitting the detail readme. thanks

Animesh Karnewar · Answer 5 · Mon Sep 03 2018 13:11:43 GMT+0800 (China Standard Time)

Hi @lovejasmine, I have been busy with my GAN projects lately. Will work on the documentation here soon.

Cheers!

Animesh Karnewar · Answer 6 · Fri Sep 21 2018 16:19:39 GMT+0800 (China Standard Time)

I have created an initial README.md. let me know if any specific details are to be added.

Cheers 🍻!
@akanimax

yuanke · Answer 7 · Fri Sep 21 2018 23:58:26 GMT+0800 (China Standard Time)

I had pass fast_prep_1,2
however, it looks missing some file when trainging:

mldl@ub1604:/ub16_prj/natural-language-summary-generation-from-structured-data/TensorFlow_implementation$ python3 trainer_with_copy_net.py
Traceback (most recent call last):
File "trainer_with_copy_net.py", line 4, in
from Summary_Generator.Model import *
File "/home/mldl/ub16_prj/natural-language-summary-generation-from-structured-data/TensorFlow_implementation/Summary_Generator/Model.py", line 16, in
from seq2seq.metrics.bleu import moses_multi_bleu
ImportError: No module named 'seq2seq.metrics'
mldl@ub1604:/ub16_prj/natural-language-summary-generation-from-structured-data/TensorFlow_implementation$ ll
total 1704668
drwxrwxr-x 4 mldl mldl 4096 9月 21 23:33 ./
drwxrwxr-x 7 mldl mldl 4096 9月 21 23:55 ../
-rw-rw-r-- 1 mldl mldl 16179 9月 21 17:14 Data_Preprocessor.ipynb
-rw-rw-r-- 1 mldl mldl 7035 9月 21 17:14 Data_Preprocessor.py
-rw-rw-r-- 1 mldl mldl 3830 9月 21 17:34 fast_data_preprocessor_part1.py
-rw-rw-r-- 1 mldl mldl 3001 9月 21 17:14 fast_data_preprocessor_part2.py
-rw-rw-r-- 1 mldl mldl 74 9月 21 17:14 .gitignore
-rw-rw-r-- 1 mldl mldl 61340 9月 21 17:14 Graph_Builder.ipynb
-rw-rw-r-- 1 mldl mldl 1995 9月 21 17:14 inferencer.py
drwxrwxr-x 2 mldl mldl 4096 9月 21 17:14 seq2seq/
drwxrwxr-x 5 mldl mldl 4096 9月 21 17:19 Summary_Generator/
-rw-rw-r-- 1 mldl mldl 1745437383 9月 21 23:36 temp.pickle
-rw-rw-r-- 1 mldl mldl 4100 9月 21 17:14 trainer_with_copy_net.py
-rw-rw-r-- 1 mldl mldl 4360 9月 21 17:14 trainer_without_copy_net.py
mldl@ub1604:/ub16_prj/natural-language-summary-generation-from-structured-data/TensorFlow_implementation$ ll seq2seq/
total 8
drwxrwxr-x 2 mldl mldl 4096 9月 21 17:14 ./
drwxrwxr-x 4 mldl mldl 4096 9月 21 23:33 ../
mldl@ub1604:/ub16_prj/natural-language-summary-generation-from-structured-data/TensorFlow_implementation$

Animesh Karnewar · Answer 8 · Sat Sep 22 2018 00:18:15 GMT+0800 (China Standard Time)

@lovejasmine, Oh yes! The seq-2-seq is just a link to the repo. Will fix it soon. Thanks for pointing.
Just a quick fix for you -> clone this repo seq2seq locally in this project and run the setup.py as follows:

(your_venv)$ python setup.py install

And, then run the training script. Thanks.

Best regards,
@akanimax

yuanke · Answer 9 · Sat Sep 22 2018 00:43:32 GMT+0800 (China Standard Time)

on py3+tf1.0.0

(.venv) mldl@ub1604:/ub16_prj/natural-language-summary-generation-from-structured-data/TensorFlow_implementation$
(.venv) mldl@ub1604:/ub16_prj/natural-language-summary-generation-from-structured-data/TensorFlow_implementation$ PYTHONPATH=seq2seq python3 trainer_with_copy_net.py
I tensorflow/stream_executor/dso_loader.cc:135] successfully opened CUDA library libcublas.so.8.0 locally
I tensorflow/stream_executor/dso_loader.cc:135] successfully opened CUDA library libcudnn.so.5 locally
I tensorflow/stream_executor/dso_loader.cc:135] successfully opened CUDA library libcufft.so.8.0 locally
I tensorflow/stream_executor/dso_loader.cc:135] successfully opened CUDA library libcuda.so.1 locally
I tensorflow/stream_executor/dso_loader.cc:135] successfully opened CUDA library libcurand.so.8.0 locally
Unpickling the data from the disc ...
Traceback (most recent call last):
File "trainer_with_copy_net.py", line 83, in
X, Y = synch_random_shuffle_non_np(zip(field_encodings, content_encodings), label_encodings)
File "/home/mldl/ub16_prj/natural-language-summary-generation-from-structured-data/TensorFlow_implementation/Summary_Generator/Tensorflow_Graph/utils.py", line 63, in synch_random_shuffle_non_np
X[:], Y[:] = zip(*combined)
TypeError: 'zip' object does not support item assignment
(.venv) mldl@ub1604:~/ub16_prj/natural-language-summary-generation-from-structured-data/TensorFlow_implementation$

Animesh Karnewar · Answer 10 · Sat Sep 22 2018 13:51:06 GMT+0800 (China Standard Time)

@lovejasmine,
Fixed in 4a164c81d88dba1eaf9c1a3737e127127ff6471a. Thanks!

Best regards,
@akanimax

Animesh Karnewar · Answer 11 · Thu Nov 22 2018 12:21:43 GMT+0800 (China Standard Time)

Closing this issue now

Thanks
👍