Skip to content

Latest commit

 

History

History
20 lines (18 loc) · 810 Bytes

README.md

File metadata and controls

20 lines (18 loc) · 810 Bytes
Raw

sentiment-clf

Text sentiment classification – wandb report.

Setup

First you should clone this repository by running

git clone https://github.com/hzchet/sentiment-clf.git
cd sentiment-clf

Then modify Makefile – change DATA_DIR (where your data is located) and SAVE_DIR (where you want to save all the logs and checkpoints) and also specify the index of your local GPU. After that build and run docker container by using the following command:

make build && make run

Training

In order to train models with your own configurations, run the following command

python3 train.py --config src/configs/my_config.json

src/configs/my_config.json – is where you should specify the model architecture and training details.