Logo Lanfrica

SameC137/SpeechToText

Domaine:

natural language processing

Type de record:

software
Créateur:
Sam
Hôte:
An end to end speech recognition system for Swahili audio # African language Speech Recognition - Speech-to-Text The World Food Program wants to deploy an intelligent form that collects nutritional information of food bought and sold at markets in two different countries in Africa - Ethiopia and Kenya. The design of this intelligent form requires selected people to install an app on their mobile phone, and whenever they buy food, they use their voice to activate the app to register the list of items they just bought in their own language. The intelligent systems in the app are expected to live to transcribe the speech-to-text and organize the information in an easy-to-process way in a database. ## Folder/File structure for branch artifacts-contains artifacts such meta files and other artifacts generated through the project notebook-contains notebooks for describing the functionality of the the classes to achieve the meta generation and the preprocessing scripts-contains scripts for Meta generation, preprocessing and feature extraction data.dvc- DVC File for versioning of the data requirements.txt- dependencies for code inside this branch ## Data Dataset for Swahili- ALFFA_PUBLIC ## Data Features Input features (X): audio clips of spoken words Target labels (y): text transcript of what was spoken ## Models Used ### Simple RNN ### Bidirectional RNN ### CNN RNN model ### Deep speech 2 layers ### Deep Speech with RNN layer Bidirectional ### Deep Speech 3 Layers