https://github.com/akensert/kaggle-tweet-sentiment-extraction
Aims to extract sentiments from paragraphs using advanced language models like XLNet and RoBERTa .
https://github.com/akensert/kaggle-tweet-sentiment-extraction
Last synced: about 1 year ago
JSON representation
Aims to extract sentiments from paragraphs using advanced language models like XLNet and RoBERTa .
- Host: GitHub
- URL: https://github.com/akensert/kaggle-tweet-sentiment-extraction
- Owner: akensert
- Created: 2020-03-24T20:22:19.000Z (about 6 years ago)
- Default Branch: master
- Last Pushed: 2023-06-12T21:28:24.000Z (about 3 years ago)
- Last Synced: 2024-12-10T09:08:38.562Z (over 1 year ago)
- Language: Python
- Homepage:
- Size: 2.72 MB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 3
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
### Tweet-Sentiment-Extraction Challenge
#### Requirements
Python >= 3.6
TensorFlow-GPU >= 2.0
To install required third-party Python packages, run `pip install -r requirements.txt`
#### Dataset
Datasets can be found [here](https://www.kaggle.com/c/tweet-sentiment-extraction/data) and should be put inside `input/tweet-sentiment-extraction/`
#### Modeling
To fit and predict with the transformer model(s), first run `chmod +x run.sh` (only has to be run once), then `FOLD=0 MODEL=xlnet ./run.sh` to run the xlnet transformer
Highest scoring model will be saved in `src/tweet-sentiment-extraction/weights/`. See `infer.py` on how to make predictions with saved model weights.