|
|
--- |
|
|
license: mit |
|
|
language: |
|
|
- en |
|
|
- mr |
|
|
tags: |
|
|
- DID |
|
|
- NLTM |
|
|
- NITGoa |
|
|
--- |
|
|
|
|
|
For more information about the model as a whole: <a href="https://github.com/NLTM-NITG/Dialect-Identification">NLTM-National Institute of Technology Dialect-Identification </a> |
|
|
|
|
|
Steps to use the model: |
|
|
1) Download the models and .py files |
|
|
2) Rename the paths to the models and to your corresponding audio file |
|
|
3) Instance a model object: <b>model = DID_Model()</b> |
|
|
4) Load the weights for the model by running: <b>model.load_weights(model_path)</b> |
|
|
5) Predict the dialect through inference: <b>model.predict_dialect(aud_path, wave2vec_model_path)</b> |
