w2v2-speaker
Research code for the paper "Fine-tuning wav2vec2 for speaker recognition" found at https://arxiv.org/abs/2109.15053
File Explorer
Download Latest Version (.zip)- debugging.yaml
- default_speech.yaml
- none.yaml
- speaker_default.yaml
- speaker_early_stopping.yaml
- speaker.yaml
- speech.yaml
- librispeech.yaml
- voxceleb1.yaml
- voxceleb1_pairs.yaml
- voxceleb1_triplets.yaml
- voxceleb2.yaml
- voxceleb2_pairs.yaml
- voxceleb2_test_everyone.yaml
- voxceleb2_test_hard.yaml
- voxceleb2_triplets.yaml
- wav2vec_base_pipeline.yaml
- wav2vec_full_seq_pipeline.yaml
- wav2vec_pair_pipeline.yaml
- wav2vec_short_seq_pipeline.yaml
- xvector_all_augment_pipeline.yaml
- xvector_dropout_augment_pipeline.yaml
- xvector_pipeline.yaml
- xvector_rirs_augment.yaml
- shards_librispeech.yaml
- shards_voxceleb.yaml
- cosine_distance.yaml
- cosine_distance_with_train_data.yaml
- lda.yaml
- plda.yaml
- speaker_dummy.yaml
- speaker_ecapa_tdnn.yaml
- speaker_wav2vec2_aam.yaml
- speaker_wav2vec2_ce.yaml
- speaker_wav2vec2_ctc.yaml
- speaker_wav2vec2_pairs.yaml
- speaker_wav2vec2_triplet.yaml
- speaker_wav2vec2_triplet_ce.yaml
- speaker_xvector.yaml
- speech_wav2vec2_ctc.yaml
- slurm.yaml
- dummy.yaml
- ecapa_tdnn.yaml
- wav2spk.yaml
- wav2vec2_fc.yaml
- wav2vec2_fc_letter.yaml
- wav2vec2_paired.yaml
- wav2vec_fc.yaml
- wav2vec_xvector.yaml
- xvector.yaml
- adam.yaml
- sgd.yaml
- aam_softmax.yaml
- binary_cross_entropy.yaml
- cross_entropy.yaml
- ctc.yaml
- triplet.yaml
- triplet_ce.yaml
- constant.yaml
- cyclic.yaml
- exp_decay.yaml
- one_cycle.yaml
- reduce_on_plateau.yaml
- schedule_wav2spk.yaml
- schedule_wav2vec_fan_etal.yaml
- tri_stage.yaml
- advanced.yaml
- simple.yaml
- lr_and_aam_loss.yaml
- lr_and_pooling.yaml
- lr_and_schedule_search.yaml
- default.yaml
- debug_trainer.yaml
- trainer.yaml
- predict.yaml
- train_eval.yaml
- config.yaml
- hydra.yaml
- overrides.yaml
- events.out.tfevents.1631794798.katara.82853.0
- data.json
- plot.png
- plot_lr_eer.png
- plot_lr_eer_zoomed.png
- run.log
- config.yaml
- hydra.yaml
- overrides.yaml
- events.out.tfevents.1631044502.katara.6664.0
- data.json
- plot.png
- plot_lr_eer.png
- plot_lr_eer_zoomed.png
- run.log
- config.yaml
- hydra.yaml
- overrides.yaml
- events.out.tfevents.1631113238.katara.16035.0
- data.json
- plot.png
- plot_lr_eer.png
- plot_lr_eer_zoomed.png
- run.log
- config.yaml
- hydra.yaml
- overrides.yaml
- events.out.tfevents.1631043151.katara.6259.0
- data.json
- plot.png
- plot_lr_eer.png
- plot_lr_eer_zoomed.png
- run.log
- config.yaml
- hydra.yaml
- overrides.yaml
- events.out.tfevents.1631793388.katara.71473.0
- data.json
- plot.png
- plot_lr_eer.png
- run.log
- config.yaml
- hydra.yaml
- overrides.yaml
- events.out.tfevents.1631794594.katara.80664.0
- data.json
- plot.png
- plot_lr_eer.png
- plot_lr_eer_zoomed.png
- run.log
- grid_search_results.csv
- plot_auto_lr.py
- plot_eer_and_lr_find.py
- plot_eer_and_lr_find_broken.py
- run_tests_pool.py
- download_and_prepare_rirs.sh
- download_librispeech.sh
- download_pretrained_models.sh
- download_voxceleb_meta.sh
- hydra_bash_complete.sh
- set_cuda_dependencies.sh
- validate_scores.py
- voxceleb2_convert_to_wav.py
- requirements_cuda101.txt
- requirements_cuda111.txt
- requirements_py1.9_cuda111.txt
- __init__.py
- input_monitor_callback.py
- memory_monitor.py
- progress_tracker_callback.py
- __init__.cpython-38.pyc
- speaker_data_module.cpython-38.pyc
- training_batch_speaker.cpython-38.pyc
- voxceleb.cpython-38.pyc
- __init__.py
- speaker_data_module.py
- training_batch_speaker.py
- voxceleb.py
- __init__.cpython-38.pyc
- librispeech.cpython-38.pyc
- speech_data_module.cpython-38.pyc
- training_batch_speech.cpython-38.pyc
- __init__.py
- librispeech.py
- speech_data_module.py
- training_batch_speech.py
- __init__.py
- __init__.py
- audio_features.py
- augment.py
- base.py
- input_normalisation.py
- random_chunks.py
- __init__.py
- collating.py
- common.py
- util.py
- __init__.py
- cosine_distance.py
- lda.py
- plda.py
- speaker_recognition_evaluator.py
- __init__.py
- wer.py
- __init__.py
- __init__.py
- embedding_masking.py
- pooling.py
- temporal_gating.py
- __init__.py
- mt_speech_speaker_module.py
- __init__.py
- dummy.py
- ecapa_tdnn.py
- paired_speaker_recognition_module.py
- speaker_recognition_module.py
- wav2spk.py
- wav2vec2_ctc.py
- wav2vec2_fc.py
- wav2vec2_paired_input.py
- wav2vec_fc.py
- wav2vec_xvector.py
- xvector.py
- __init__.py
- speech_recognition_module.py
- wav2vec2_fc_letter.py
- __init__.py
- base_lightning_module.py
- __init__.py
- wav2vec.py
- wav2vec2.py
- __init__.py
- aam_softmax.py
- binary_cross_entropy.py
- cross_entropy.py
- ctc_loss.py
- triplet_ce_loss.py
- triplet_loss.py
- __init__.py
- tri_stage.py
- __init__.py
- __init__.py
- base.py
- tokenizer_wav2vec2.py
- __init__.py
- config_util.py
- eval_metrics.py
- hydra_resolvers.py
- main.py
- predict.py
- util.py
- .env.example
- .gitignore
- convert_voxceleb2.sh
- LICENCE
- predict.py
- pyproject.toml
- README.md
- run.py
# Installation Guide
1. Get the code
git clone https://github.com/nikvaessen/w2v2-speaker
Downloads the entire project code from GitHub to your computer.
cd w2v2-speaker
Moves into the project folder you just downloaded.
2. Python
Easy RecommendedPrerequisites
$ poetry shell # enter project-specific virtual environment
Type this command into your terminal and run it.
(xxx) $ pip install -U pip
Installs the Python libraries listed in requirements.txt (or similar).
(xxx) $ poetry update # install dependencies
Type this command into your terminal and run it.
(xxx) $ pip install -r requirements/requirements_cuda101.txt # if CUDA 10.1
Installs the Python libraries listed in requirements.txt (or similar).
(xxx) $ pip install -r requirements/requirements_cuda110.txt # if CUDA 11.0
Installs the Python libraries listed in requirements.txt (or similar).
If it runs without errors and prints output in the terminal, it worked.
Pulled directly from this repo's README.
// repository documentation
Was this content helpful?
(0 ratings)