speechbrain.utils
Package containing various tools (accuracy, checkpoints …)
Calculate accuracy. |
|
Calculates Diarization Error Rate (DER) which is the sum of Missed Speaker (MS), False Alarm (FA), and Speaker Error Rate (SER) using md-eval-22.pl from NIST RT Evaluation. |
|
Library for computing the BLEU score |
|
Chaining together callables, if some require relative lengths |
|
This module implements a checkpoint saver and loader. |
|
A pipeline for data transformations. |
|
This library gathers utilities for data io operation. |
|
A dependency graph for finding evaluation order. |
|
Guard for running certain operations on main process only |
|
Edit distance and WER computation. |
|
Implements a checkpointable epoch counter (loop), optionally integrating early stopping. |
|
Utilities for hparams files |
|
Utilities for hyperparameter optimization. |
|
Managing the logger, utilities |
|
The |
|
Parallel processing tools to help speed up certain tasks like data preprocessing. |
|
Convenience functions for the simplest parameter transfer cases. |
|
Polymorphic decorators to handle PyTorch profiling and benchmarking. |
|
Superpowers which should be sparingly used. |
|
Library for checking the torchaudio backend. |
|
Loggers for experiment monitoring. |