Simba bridges the digital divide with a unified suite for African AI: the largest open-source speech benchmark and models covering 61 languages
AI & ML interests
None defined yet.
Recent Activity
A collection of all the resources that we built for the NileChat LLM project.
ARBERT and MARBERT, two powerful Transformer-based language models for Arabic
Arabic Distil Whisper Models
PEARL: A Multimodal Culturally-Aware Arabic Instruction Dataset
Models and dataset collections for our uDistilWhisper paper which was accepted at NAAC'25 main.
-
UBC-NLP/distil-large-v2-init-16-16-100K-225-sonar_sim
Automatic Speech Recognition • 0.8B • Updated -
UBC-NLP/distil-large-v2-init-16-16-100K-225-sm4t
Automatic Speech Recognition • 0.8B • Updated -
UBC-NLP/distil-large-v2-init-32-16-100K-225-sonar_sim
Automatic Speech Recognition • 1B • Updated • 2 -
UBC-NLP/distil-large-v2-init-32-16-100K-225-sm4t
Automatic Speech Recognition • 1B • Updated • 1
AraT5: Text-to-Text Transformers for Arabic Language Generation
Simba bridges the digital divide with a unified suite for African AI: the largest open-source speech benchmark and models covering 61 languages
PEARL: A Multimodal Culturally-Aware Arabic Instruction Dataset
A collection of all the resources that we built for the NileChat LLM project.
Models and dataset collections for our uDistilWhisper paper which was accepted at NAAC'25 main.
-
UBC-NLP/distil-large-v2-init-16-16-100K-225-sonar_sim
Automatic Speech Recognition • 0.8B • Updated -
UBC-NLP/distil-large-v2-init-16-16-100K-225-sm4t
Automatic Speech Recognition • 0.8B • Updated -
UBC-NLP/distil-large-v2-init-32-16-100K-225-sonar_sim
Automatic Speech Recognition • 1B • Updated • 2 -
UBC-NLP/distil-large-v2-init-32-16-100K-225-sm4t
Automatic Speech Recognition • 1B • Updated • 1
ARBERT and MARBERT, two powerful Transformer-based language models for Arabic
AraT5: Text-to-Text Transformers for Arabic Language Generation
Arabic Distil Whisper Models