nouhadziri / THRED

The implementation of the paper "Augmenting Neural Response Generation with Context-Aware Topical Attention"
https://arxiv.org/abs/1811.01063
MIT License
111 stars 25 forks source link

python main.py --mode train #1

Closed zysNLP closed 5 years ago

zysNLP commented 5 years ago

Excuse me, in the following command: python main.py --mode train --config \ --train_data --dev_data --test_data \ --model_dir

Does YAML_FILE means .yml file as thred_medium.yml? But where are the TRAIN_DATA,DEV_DATA and TEST_DATA? Does it means I should download all the .bz2 files you mentioned in THRED/corpora/reddit/README.rd——“Reddit monthly data needs to be downloaded from here for comments and here for submissions”?

nouhadziri commented 5 years ago

Thanks for you interest in our work!

Any Yaml file (including the thred_medium.yml) in the conf folder would work. It depends on what model you want to train.

We've had some difficulties to publish the data. However, the issues are mostly resolved and we will publish the data soon. In the mean time, you can build the data from scratch based on the doc provided here.

zysNLP commented 5 years ago

Ok,thank you very much.Expect publish data sets.