IBM Releases Granite Speech 5.0 2B Model
IBM has released Granite Speech 5.0 2B, a new speech-language model designed for enterprise speech tasks alongside updates to other models in the collection. This model is part of a broader update that includes revised versions like granite-speech-5.0-470m-turboctc and granite-speech-4.1-2b-plus. This release provides enterprises with robust open-source tools for high-accuracy speech transcription and translation. It expands local AI capabilities for developers and organizations looking to process audio data efficiently. The Granite Speech model is supported natively in transformers version 4.52.1 and higher. It covers a full range of enterprise speech workflows including both transcription and translation.
## BACKGROUND
Speech-language models are advanced AI systems that bridge audio processing and language understanding, allowing machines to convert spoken words into text or translate them directly. IBM's Granite collection comprises open foundation models tailored for enterprise applications across various modalities.