Instructions to use czyhust/finetune_moss-sortformer with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- NeMo
How to use czyhust/finetune_moss-sortformer with NeMo:
# tag did not correspond to a valid NeMo domain.
- Notebooks
- Google Colab
- Kaggle
File size: 1,038 Bytes
74e4281 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 | . ./path.sh
set -euo pipefail
train_manifest_path=data/dump/synthetic_run-debug-train/train.json
dev_manifest_path=${train_manifest_path}
exp_dir=./exp/streaming_sortformer_diar_train
exp_name=sample_train/debug
config_name=streaming_sortformer_diarizer_4spk-v2.yaml
. ./src/finetune_pipeline/scripts/parse_options.sh || exit 1
config_path='conf' # relative path w.r.t. train script
train_script=./src/finetune_pipeline/bin/streaming_sortformer_diar_train.py
echo "config_path: $(dirname ${train_script})/${config_path}/${config_name}"
echo "train_manifest_path: ${train_manifest_path}"
echo "dev_manifest_path: ${dev_manifest_path}"
echo "exp_dir: ${exp_dir}/${exp_name}"
python ${train_script} \
--config-path="${config_path}" \
--config-name="${config_name}" \
trainer.precision='bf16' \
trainer.devices=1 \
model.train_ds.manifest_filepath="${train_manifest_path}" \
model.validation_ds.manifest_filepath="${dev_manifest_path}" \
exp_manager.name="${exp_name}" \
exp_manager.exp_dir="${exp_dir}" |