`pip install -r requirements.txt` causes an error

[forest1988]

Thank you for sharing your great work in GitHub! I'm working on NLP, especially on story analysis and generation, so I'm very interested in your proposed metric "UNION".

I'd like to use the metric, but have trouble installing prerequisites. If you know how to avoid the error, could you please tell me how to solve it?

First, I tried pip install -r requirements.txt in Python 3.8 environment, but failed in installing Tensorflow v1. Then, I downgraded Python to 3.7 (as you indicated in "Prerequisites" in "README.md"), and TensorFlow v1 could be installed. However, now I come across another error saying the version of regex "Could not find".

(If you allow me to add information, "tensorflow-gpu==1.14.0" seems duplicated in requirements.txt. Is it something you intended?)

<username>@<servername> $ conda activate py37_union
(py37_union) <username>@<servername> $ pwd
/********/workspace/Clone/UNION
(py37_union) <username>@<servername> $ pip install -r requirements.txt
Collecting tensorflow-gpu==1.14.0
  Downloading tensorflow_gpu-1.14.0-cp37-cp37m-manylinux1_x86_64.whl (377.1 MB)
     |████████████████████████████████| 377.1 MB 5.2 kB/s
Collecting numpy==1.18.1
  Downloading numpy-1.18.1-cp37-cp37m-manylinux1_x86_64.whl (20.1 MB)
     |████████████████████████████████| 20.1 MB 38.1 MB/s
ERROR: Could not find a version that satisfies the requirement regex==2.5.76 (from -r requirements.txt (line 4)) (from versions: 2013-02-16, 2013-02-23, 2013-03-11, 2013-05-21, 2013-06-05, 2013-06-26, 2013-08-04, 2013-10-04, 2013-10-12, 2013-10-21, 2013-10-22, 2013-10-23, 2013-10-24, 2013-10-25, 2013-10-26, 2013-11-29, 2013-12-31, 0.1.20100217, 0.1.20100226, 0.1.20100305, 0.1.20100323, 0.1.20100331, 0.1.20100706, 0.1.20100706.1, 0.1.20100709, 0.1.20100709.1, 0.1.20100719, 0.1.20100725, 0.1.20100814, 0.1.20100816, 0.1.20100824, 0.1.20100912, 0.1.20100913, 0.1.20100918, 0.1.20101009, 0.1.20101029, 0.1.20101030b0, 0.1.20101030, 0.1.20101101, 0.1.20101102a0, 0.1.20101102, 0.1.20101106, 0.1.20101113, 0.1.20101120, 0.1.20101121, 0.1.20101123, 0.1.20101130, 0.1.20101207, 0.1.20101210, 0.1.20101224, 0.1.20101228a0, 0.1.20101228, 0.1.20101229, 0.1.20101230, 0.1.20101231, 0.1.20110104, 0.1.20110106, 0.1.20110124, 0.1.20110313, 0.1.20110314, 0.1.20110315, 0.1.20110429, 0.1.20110502, 0.1.20110504, 0.1.20110510, 0.1.20110514, 0.1.20110524, 0.1.20110608a0, 0.1.20110608, 0.1.20110609, 0.1.20110610, 0.1.20110616, 0.1.20110623a0, 0.1.20110623, 0.1.20110627, 0.1.20110702, 0.1.20110717, 0.1.20110917a0, 0.1.20110917, 0.1.20110922a0, 0.1.20110922, 0.1.20110927, 0.1.20110929, 0.1.20111004, 0.1.20111005, 0.1.20111006, 0.1.20111014, 0.1.20111103, 0.1.20111223, 0.1.20120103, 0.1.20120105, 0.1.20120112, 0.1.20120114, 0.1.20120115, 0.1.20120119, 0.1.20120122, 0.1.20120123, 0.1.20120126, 0.1.20120128, 0.1.20120129, 0.1.20120208, 0.1.20120209, 0.1.20120301, 0.1.20120303, 0.1.20120316, 0.1.20120317, 0.1.20120323, 0.1.20120416, 0.1.20120502, 0.1.20120503, 0.1.20120504, 0.1.20120506, 0.1.20120611, 0.1.20120613, 0.1.20120705, 0.1.20120708, 0.1.20120709, 0.1.20120710, 0.1.20120803, 0.1.20120825, 0.1.20120904, 0.1.20121008, 0.1.20121017, 0.1.20121031, 0.1.20121105, 0.1.20121113, 0.1.20121120, 0.1.20121216, 0.1.20130120, 0.1.20130124, 0.1.20130125, 2014.1.10, 2014.1.20, 2014.1.30, 2014.2.16, 2014.2.19, 2014.4.10, 2014.5.17, 2014.5.23, 2014.6.28, 2014.8.15, 2014.8.28, 2014.9.18, 2014.9.22, 2014.10.1, 2014.10.2, 2014.10.7, 2014.10.9, 2014.10.23, 2014.10.24, 2014.11.3, 2014.11.13, 2014.11.14, 2014.12.15, 2014.12.24, 2015.3.18, 2015.5.7, 2015.5.10, 2015.5.28, 2015.6.2, 2015.6.4, 2015.6.9, 2015.6.10, 2015.6.14, 2015.6.15, 2015.6.19, 2015.6.21, 2015.6.24, 2015.7.12, 2015.7.19, 2015.9.14, 2015.9.15, 2015.9.23, 2015.9.28, 2015.10.1, 2015.10.5, 2015.10.22, 2015.10.29, 2015.11.5b0, 2015.11.7, 2015.11.8, 2015.11.9, 2015.11.12, 2015.11.14, 2015.11.22, 2016.1.10, 2016.2.23, 2016.2.24, 2016.2.25, 2016.3.2, 2016.3.24, 2016.3.26, 2016.3.31, 2016.4.1, 2016.4.2, 2016.4.3, 2016.4.8, 2016.4.15, 2016.4.25, 2016.5.13, 2016.5.14, 2016.5.15, 2016.5.23, 2016.6.2, 2016.6.5, 2016.6.14, 2016.6.19, 2016.6.24, 2016.7.14, 2016.7.21, 2016.8.27, 2016.9.22, 2016.10.22, 2016.11.18, 2016.11.21, 2016.12.27, 2017.1.12, 2017.1.14, 2017.1.17, 2017.2.8, 2017.4.5, 2017.4.23, 2017.4.29, 2017.5.26, 2017.6.7, 2017.6.20, 2017.6.23, 2017.7.11, 2017.7.26, 2017.7.28, 2017.9.23, 2017.11.8, 2017.11.9, 2017.12.5, 2017.12.9, 2017.12.12, 2018.1.10, 2018.2.3, 2018.2.8, 2018.2.21, 2018.6.6, 2018.6.9, 2018.6.20, 2018.6.21, 2018.7.11, 2018.8.17, 2018.8.29, 2018.11.2, 2018.11.3, 2018.11.6, 2018.11.7, 2018.11.22, 2019.1.23, 2019.1.24, 2019.2.3, 2019.2.5, 2019.2.6, 2019.2.7, 2019.2.18, 2019.2.19, 2019.2.20, 2019.2.21, 2019.3.8, 2019.3.9, 2019.3.12, 2019.4.9, 2019.4.10, 2019.4.12, 2019.4.14, 2019.5.25, 2019.6.2, 2019.6.5, 2019.6.8, 2019.8.19, 2019.11.1, 2019.12.9, 2019.12.17, 2019.12.18, 2019.12.19, 2019.12.20, 2020.1.7, 2020.1.8, 2020.2.18, 2020.2.20, 2020.4.4, 2020.5.7, 2020.5.13, 2020.5.14, 2020.6.7, 2020.6.8, 2020.7.14, 2020.9.27, 2020.10.11)
ERROR: No matching distribution found for regex==2.5.76 (from -r requirements.txt (line 4))

Thank you in advance!

[JianGuanTHU]

Thank you for pointing out the bug! I have fixed the requirements.txt. Hopefully, it will work for you.

[forest1988]

Thank you for your quick work and response! I'll try it!

[forest1988]

Thanks for your support, I could run the following steps!

1. Constructing Negative Samples
2. Training of UNION

It seems step 1 needs one more prerequisite, "nltk". At step 2, I used the 10 samples in GitHub (it is less than 1000 steps, so no model is saved. Is my understanding right?).

After this, I'd like to try the full data and the fine-tuned models you kindly provide via cloud storage!

Thanks again!

[JianGuanTHU]

Exactly! Thank you very much for your comments.