aalto-speech Goto Github PK

ftk

This toolkit contains programs for segmenting strings and training string segmentation models. It has been developed primarily for learning units for speech recognition, but can be used for other purposes as well.

interspeech2019_karhila_et_al

Compendium for the paper "Transparent pronunciation scoring using articulatorily weighted phoneme edit distance" by Karhila, Smolander, Ylinen & Kurimo submitted to Interspeech 2019

ite-typing-dataset

Scripts and jupyter notebooks to process and analyse ITE typing dataset

kaldi-sb-north-sme

Kaldi + SpeechBrain + W2V2 models for Northern Sami

kaldi-utensils

Kaldi utility scripts

l2-game

Speech recognition based game for second language learning

l2-speech-scoring-tools

Implementation of automatic speech rating systems for second language (L2) learners of Finnish and Finland Swedish

lahjoita-puhetta-baseline-kaldi

Kaldi ASR system for Lahjoita puhetta corpus

lahjoita-puhetta-baseline-wav2vec2

Baseline self-supervised Wav2Vec2 ASR system for Lahjoita puhetta corpus

lahjoita-puhetta-metadata-classification

lahjoita-puhetta-resources

A collection of resources related to the Lahjoita puhetta speech corpus.

modules

Installation scripts for used modules in Aalto ASR research group

moodle-mod_digitala

DigiTala is a Moodle plugin for assessing L2 Finnish and Swedish speech automatically. Cite as: "von Zansen, A., Alanen, T., Al-Ghezi, R., Erkkilä, J., Harjunpää, T., Heijala, M., Kallio, H. (2022). DigiTala Moodle plugin. https://github.com/aalto-speech/moodle-mod_digitala "

morfessor

Morfessor is a tool for unsupervised and semi-supervised morphological segmentation

morfessor-demo

Morfessor demonstration

morfessor-emprune

Morfessor EM+Prune

morphological-classes

Morphological word class training

multitask-wav2vec2

Custom 🤗 Transformers for training multi-task wav2vec2 models that perform ASR and speech classification tasks simultaneously as described in Getman, Y., Al-Ghezi, R., Grósz, T., Kurimo, M. (2023) Multi-task wav2vec2 Serving as a Pronunciation Training System for Children.

ner-asr

Named Entity Recognition for Finnish Language

neural-entity-linking-for-finnish

A Neural entity linking system trained on a Finnish sports magazine data

northern-sami-asr

Scripts for adapting large speech foundation models for Northern Sámi ASR

paragui

cross-platform widgetset for SDL (a new home for old code)

rl-klm

RL-KLM implementation that can be used to estimate task completion times for user interface.

run-nemo-on-puhti

This directory runs NeMo on Puhti

say-it-again-kid-pronunciation-learning

Privacy policies for the language learning games developed in collaboration with University of Helsinki Cognitive Brain Research group.

sb-2015-2020-kevat_e2e

AED implementations for Finnish parliament Train20 (Includes Train16 and Train Comb as well)

sb-fin-parl-2015-2020-kevat

SpeechBrain recipes for Finnish Parliament data - HMM/DNN

sb-fin-parl-models

SpeechBrain baseline recipes for Finnish Parliament data

sb-libri-hmmdnn

Librispeech HMM/DNN and AED SpeechBrain experiments

setup-asr-on-csc

How to setup a Kaldi and SpeechBrain environment on CSC Puhti