Source
Extremely Fast and Accurate Transcription with Granite Speech 5.0 Turbo CTC
IBM released two compact 470M-parameter English speech recognition models: granite-speech-5.0-470m-turboctc and granite-speech-5.0-470m-turboctc-nc.
Why it matters
Their small memory footprint and high speed make them suitable for performing speech-to-text tasks on edge devices with limited computing power.
The details
- Can transcribe over 3.5 hours of speech per second on NVIDIA H200 GPUs.
- One model is Apache 2.0 licensed, while the other is noncommercial CC-BY-NC-SA-4.0.
- The encoder-only design provides 20x faster throughput than previous Granite Speech models.
Get the weekly recap
The stories like this one, picked and explained — once a week, straight to your inbox.