You may see the report if you follow this link

Firstly, install needed requirements for running model

pip install -r ./requirements.txt

Use bash script to download trained model

cd ./default_test_model
./download.sh

It will be placed to ./default_test_model/checkpoint.pth

If you have some issues using bash utilities, you may download model directly from google drive

You may check this model on your data. Put your audio to test folder test_data/audio.

python test.py \
   -r default_test_model/checkpoint.pth \
   -t test_data \
   -o test_result.json

Or you can check EER and predictions on ASVDataset. Then provide config link to arguments

python test.py \
   -r default_test_model/checkpoint.pth \
   -c default_test_config.json \
   -o test_result.json

This repository is based on a heavily modified fork of pytorch-template repository.

You can use this project with docker. Quick start:

docker build -t my_src_image .
docker run \
   --gpus '"device=0"' \
   -it --rm \
   -v /path/to/local/storage/dir:/repos/asr_project_template/data/datasets \
   -e WANDB_API_KEY=<your_wandb_api_key> \
	my_src_image python -m unittest

Notes:

• -v /out/of/container/path:/inside/container/path -- bind mount a path, so you wouldn't have to download datasets at the start of every docker run.
• -e WANDB_API_KEY=<your_wandb_api_key> -- set envvar for wandb (if you want to use it). You can find your API key here: https://wandb.ai/authorize