45 lines
1.5 KiB
Markdown
45 lines
1.5 KiB
Markdown
## Requirements
|
|
Python version - 2.7
|
|
|
|
You need to download Google news vectors 300 dimensional bin file and place
|
|
it in the main directory to run the code
|
|
You will need to have Theano=0.7 to run the code. Please refer : https://github.com/nestle1993/SE16-Task6-Stance-Detection
|
|
|
|
## Usage
|
|
#### Training and Testing
|
|
python run.py <Path to the Data Directory>
|
|
#### Printing the scores for each dataset within data directory
|
|
python score.py <Path to the Data Directory>
|
|
#### Printing the tuned hyperparameters for each dataset within data directory
|
|
python get_params.py <Path to the Data Directory>
|
|
|
|
**Note:** The Data Directory be one of the preprocessed folders created by the preprocessing script.
|
|
Example directory structure of the data directory for SemEval Data:
|
|
|
|
Data_SemE_P
|
|
├── AT
|
|
| ├── test_clean.txt
|
|
| ├── test_preprocessed.csv
|
|
| ├── train_clean.txt
|
|
| └── train_preprocessed.csv
|
|
├── CC
|
|
| ├── test_clean.txt
|
|
| ├── test_preprocessed.csv
|
|
| ├── train_clean.txt
|
|
| └── train_preprocessed.csv
|
|
├── FM
|
|
| ├── test_clean.txt
|
|
| ├── test_preprocessed.csv
|
|
| ├── train_clean.txt
|
|
| └── train_preprocessed.csv
|
|
├── HC
|
|
| ├── test_clean.txt
|
|
| ├── test_preprocessed.csv
|
|
| ├── train_clean.txt
|
|
| └── train_preprocessed.csv
|
|
└── LA
|
|
├── test_clean.txt
|
|
├── test_preprocessed.csv
|
|
├── train_clean.txt
|
|
└── train_preprocessed.csv
|