search

adf1178 / QueCos_code

The code repo for <Enriching Query Semantics for Code Search with Reinforcement Learning>
3 stars 0 forks source link

Detailed documentation and data #1

Open hxue3 opened 2 years ago

hxue3 commented 2 years ago

Hi, I am trying to use your code. However, I noticed this repo is not a compeleted version, as training data is missing.

Is there a more detailed dcoumentation how to use your code? And if you can share the training dataset as well.

adf1178 commented 2 years ago

Hi~ Sorry for the unclear documentation of the code ... the implementation of QueCos is very chaotic but I think the code is completed. The main part of QueCos is in src/ke, where it connects different code bases UNIF, DeepCS, OCoR and CodeBERT. Firstly, you can train a code search model such as UNIF and put the trained model in the corresponding dir. Then modify the argument --sent_reward to use pre-trained UNIF model to compute the reward. In src/ke/lib/metric/RetReward.py we implement the reward computation function for four code search models. As for data, I am so sorry there is something wrong when we crawled the data such as repetition :( if you want training data to run the script, we can send it to you.

Ifra25 commented 2 years ago

Hi, can you please share the training data also.

xiangzheng666 commented 2 years ago

Hi, can you please share the training data also.

SIPUNK commented 1 year ago

Hi, can you please share the training data also.

chengong12 commented 10 months ago

Hi, can you please share the training data also.