diff --git a/README.md b/README.md index 8c66bc45957c41393069d35b01bb5edad5cb2d54..d083edf0a2bfed574f2bf4491a65ef4fade96fa5 100644 --- a/README.md +++ b/README.md @@ -4,12 +4,12 @@ U-NET based feature extractor for text-independent speaker verification. This re REQUIREMENTS: ----------------------------------------------------- -Python 2.7 -PyTorch from http://pytorch.org/ --- This project was implemented with version 0.3.0.post4 -librosa -scipy -numpy +..* Python 2.7 +..* PyTorch from http://pytorch.org/ +....- This project was implemented with version 0.3.0.post4 +..* librosa +..* scipy +..* numpy USAGE: ----------------------------------------------------- @@ -77,7 +77,7 @@ convert_pickle_to_kaldi_ark.py -i spec_BNF -if ${OUT_LIST_FOLDER}/voxceleb1_test convert_pickle_to_kaldi_ark.py -i spkr_BNF -if ${OUT_LIST_FOLDER}/voxceleb1_test/vad.scp -o spkr_feats_test/raw_embedding -of ${OUT_LIST_FOLDER}/voxceleb1_test/spkr_feats.scp ``` -Before running the i-vector system with spec BNFs change the ```fests.scp``` with: +Before running the i-vector system with spec BNFs change the ```feats.scp``` with: ``` mv ${OUT_LIST_FOLDER}/voxceleb1_train/feats.scp ${OUT_LIST_FOLDER}/voxceleb1_train/mfcc_feats.scp mv ${OUT_LIST_FOLDER}/voxceleb1_test/feats.scp ${OUT_LIST_FOLDER}/voxceleb1_test/mfcc_feats.scp @@ -89,7 +89,7 @@ Now change the data/train to data/voxceleb1_train in ```${KALDI_ROOT}/egs/voxcel -Similarly Before running the i-vector system with spkr BNFs change the ```fests.scp``` with: +Similarly Before running the i-vector system with spkr BNFs change the ```feats.scp``` with: ``` mv ${OUT_LIST_FOLDER}/voxceleb1_train/feats.scp ${OUT_LIST_FOLDER}/voxceleb1_train/spec_feats.scp mv ${OUT_LIST_FOLDER}/voxceleb1_test/feats.scp ${OUT_LIST_FOLDER}/voxceleb1_test/spec_feats.scp