| --- |
| license: mit |
| language: |
| - en |
| tags: |
| - spoken-question-answering |
| - speech |
| - question-answering |
| - checkpoints |
| - reproducibility |
| base_model: mistralai/Mistral-7B-Instruct-v0.2 |
| --- |
| # SpokenLLM |
|
|
| This repository contains the model checkpoints for my master's thesis on Spoken Question Answering. |
|
|
| The code for training, inference, and evaluation is available in the corresponding GitHub repository. The checkpoints in this Hugging Face repository are provided for reproducibility. For the reported evaluation results in the thesis, the epoch-10 checkpoints were used. |
|
|
| ## Checkpoints |
|
|
| All trained checkpoints are uploaded in the `checkpoints/` directory. The main evaluation results reported in the thesis are based on the `epoch_10` checkpoints. |
|
|
| ## Notes |
|
|
| These checkpoints are intended for academic reproducibility of the thesis experiments. |
