Skip to content
/ senh Public

Speech enhancement package to create datasets, evaluate models, train models.

Notifications You must be signed in to change notification settings

CanKeles5/senh

Repository files navigation

senh

Speech enhancement package to create datasets, evaluate models, train models (To-do).

Main features:

  • Create custom datasets with specific SNR rates and reverb level
  • Evaulate models for different cases such as clean noise, reverbarant noise, speaker with specific background SNR
  • Train models on custom dataset, options to fine tune models from HuggingFace. Currently supporting ConvTasNet and DPTNet

Dataset for model evaluation can be found here: https://huggingface.co/datasets/cankeles/eval-senh

About

Speech enhancement package to create datasets, evaluate models, train models.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published