Clone this repo and run
cd ADL-Final
apt-get update && apt-get install -y libsndfile1 ffmpeg
bash ./reinstall.sh
To generate predictions from test dataset, run
python test.py \
model.dataset.data_dir=<data_dir_with_dialogue_data> \
split=<dev/test_seen/test_unseen>
The output files would be located in folder predictions.
The structure of the data directory should be like:
data_dir/
schemas.json
train/
dev/
test_seen/
test_unseen/
The files that are modified from the original NeMo repo to generate predictions are listed below.
NeMo/nemo/collections/nlp/data/dialogue_state_tracking/sgd/
prediction_utils.py
dataset.py
NeMo/nemo/collections/nlp/models/dialogue_state_tracking/
sgdqa_model.py