Skip to content

Verifying the identity of a person from characteristics of the voice independent from language via NVIDIA NeMo models (ECAPA-TDNN, SpeakerNet, TitaNet-L).

Notifications You must be signed in to change notification settings

Rumeysakeskin/Speaker-Verification

Repository files navigation

Speaker Verification

Speaker verification is verifying the identity of a person from characteristics of the voice independent from language via NVIDIA NeMo.

This reporisitory presents three NeMo speaker verification models:

Download Models - Save and Load Speaker Vectors

You can download Nemo models and speaker vectors for SpeakerNet, TitaNet-L, ECAPA-TDNN from files/.

Prediction

The cosine similarity metric was used for prediction.

from sklearn.metrics.pairwise import cosine_similarity
sims = cosine_similarity([vector[0]], speakers_vectors)[0]

To predict most similar speaker in test_voices/ refered to ref_voices run the following command:

python speaker_verification.py

Inference Benchmark

  • Torch Model
  • Onnx Model Run the script below to compare the inference time of SpeakerNet model.
speaker_verification_with_torch_and_onnx.ipynb

About

Verifying the identity of a person from characteristics of the voice independent from language via NVIDIA NeMo models (ECAPA-TDNN, SpeakerNet, TitaNet-L).

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published