This is a pytorch implementation of multiclass sentiment analysis on SST dataset. In here, both CNN and RNN model implementation is shown here.
| File Name | Description |
|---|---|
| model.py | Classification Models |
| utils.py | Loading dataset, model fit, model test and visualize |
| train.py | Training and testing models |
| sample_execution.ipynb | sample execution |
| run.sh | Bash file for running this program in 6 configuration |
This repository can be donwloaded as
git clone https://github.com/aminul-huq/CNN-RNN-for-Multiclass-Classification-on-SST-dataset.git
bash run.sh
Tested under python 3.6 and pytorch 1.0
| Command | Default_value |
|---|---|
| --model | 0 |
| --lr | 0.001 |
| --itr | 5 |
| --dropout | 0.5 |
| --hidden_dim | 256 |
| --device | "cpu" |
| --n_filters | 100 |
| --filter_size | 3 |