Logo Lanfrica

Kenyan Swahili ASR (clean, Nemotron-ready)

Domain:

natural language processing

Record type:

dataset
Creator:
Ton
Host:
A cleaned, validated Kenyan-Swahili ASR corpus prepared for fine-tuning streaming ASR models (e.g. NVIDIA Nemotron 3.5 ASR). This is an initial ~30h subset for pipeline validation; a larger version will follow. Audio: 16 kHz mono WAV (embedded). text: cased + punctuated transcript. target_lang: sw-KE · source, dialect, duration columns included. Source & license

Tasks

automatic speech recognitionspeech processing

Languages

Swahili

Tags

swahilikenyan-swahilispeechasr

Licenses

cc-by-4.0

Similar

afyfbadreddine77/darija-asr-clean

From a Multilingual Streaming ASR Backbone to Kenyan-Language Systems: Data-Centric Adaptation of Nemotron 3.5 for Kikuyu, Dholuo, and Kalenjin

Automatic speech recognition (ASR) for African languages is constrained by orthographic inconsistenc

Panga-Azazia/mb-TTS-ASR-ready-2

Swahili Small Clean Corpus (Demo)

One-line summaryA small high-quality Swahili text corpus (≈58k documents, ~52.8 MB) extracted from S

Swahili-ASR

Swahili-ASR is a 600-hour conversational Swahili speech corpus with fully human-generated verbatim t

ElvisGitau/swahili-ASR