Granite Speech 5.0: Transcribe, Translate, and Attribute more Accurately and Faster
• 27
AI & ML interests
Enterprise-grade AI models
Recent Activity
View all activity
Papers
ChartNet: A Million-Scale, High-Quality Multimodal Dataset for Robust Chart Understanding
SmolDocling: An ultra-compact vision-language model for end-to-end multi-modal document conversion
Articles
Efficient reasoning and thinking language models for multilingual generation, coding, and AI assistant workflows.
Efficient language models for multilingual generation, coding, RAG, and AI assistant workflows.
Ultra-compact language models designed for the edge and on-device deployment.
-
ibm-granite/granite-4.0-h-1b
Text Generation • 1B • Updated • 3.19k • 148 -
ibm-granite/granite-4.0-h-1b-base
Text Generation • 1B • Updated • 1.01k • 35 -
ibm-granite/granite-4.0-h-350m
Text Generation • 0.3B • Updated • 3.83k • 112 -
ibm-granite/granite-4.0-h-350m-base
Text Generation • 0.3B • Updated • 914 • 34
Multilingual ASR and speech-to-text (STT) models for enterprise transcription and translation.
-
ibm-granite/granite-speech-5.0-2b
Automatic Speech Recognition • 2B • Updated • 4k • 53 -
ibm-granite/granite-speech-5.0-470m-turboctc
Automatic Speech Recognition • 0.5B • Updated • 61.5k • 102 -
ibm-granite/granite-speech-5.0-470m-turboctc-nc
Automatic Speech Recognition • 0.5B • Updated • 38k • 38 -
Granite Speech Streaming Webgpu
🌍19Granite Speech 5.0 Turbo CTC Streaming Audio Demo
Multimodal models built for visual document analysis and image understanding.
-
Multimodal RAG with Granite Vision
🚀42RAG example using Granite [vision, embedding, instruct]
-
ibm-granite/granite-vision-4.1-4b
Image-Text-to-Text • 4B • Updated • 126k • 112 -
ibm-granite/granite-4.0-3b-vision
Image-Text-to-Text • 4B • Updated • 69k • 112 -
ibm-granite/granite-vision-3.3-2b
Image-to-Text • 3B • Updated • 179k • 88
Time series models for forecasting, anomaly detection, classification, and more.
-
ibm-granite/granite-timeseries-ensemble-r1
Time Series Forecasting • Updated • 320 • 8 -
ibm-granite/granite-timeseries-ttm-r3
Time Series Forecasting • 1.41M • Updated • 397k • 20 -
Tiny Time Mixers (TTMs): Fast Pre-trained Models for Enhanced Zero/Few-Shot Forecasting of Multivariate Time Series
Paper • 2401.03955 • Published • 15 -
ibm-granite/granite-timeseries-patchtst-fm-r2
Time Series Forecasting • 0.4B • Updated • 212k • 17
Language models with improved reasoning and instruction-following capabilities.
-
ibm-granite/granite-3.3-8b-instruct
Text Generation • 8B • Updated • 73.3k • 161 -
ibm-granite/granite-3.3-2b-instruct
Text Generation • 3B • Updated • 338k • 87 -
ibm-granite/granite-3.3-8b-base
Text Generation • 8B • Updated • 5.48k • 27 -
ibm-granite/granite-3.3-2b-base
Text Generation • 3B • Updated • 4.76k • 23
Artifacts that are related to curating and evaluating datasets used for Granite models.
Quantized versions of IBM Granite models.
Modular Granite models with built-in adapters, switched on as needed from a single checkpoint.
Efficient language models for multilingual generation, coding, RAG, and AI assistant workflows.
-
ibm-granite/granite-4.0-micro
Text Generation • 3B • Updated • 25.6k • 274 -
ibm-granite/granite-4.0-micro-base
Text Generation • 3B • Updated • 2.23k • 41 -
ibm-granite/granite-4.0-h-micro
Text Generation • 3B • Updated • 8.7k • 147 -
ibm-granite/granite-4.0-h-micro-base
Text Generation • 3B • Updated • 908 • 35
Embedding models (bi‑encoders and rerankers) for RAG, semantic search, and retrieval tasks.
-
ibm-granite/granite-embedding-311m-multilingual-r2
Feature Extraction • 0.3B • Updated • 98.1k • • 141 -
ibm-granite/granite-embedding-97m-multilingual-r2
Feature Extraction • 97.4M • Updated • 109k • • 147 -
ibm-granite/granite-embedding-english-r2
Feature Extraction • 0.1B • Updated • 47.2k • 92 -
ibm-granite/granite-embedding-small-english-r2
Feature Extraction • 47.7M • Updated • 6.39M • 112
Models for parsing complex PDFs and structured documents, designed to complement Docling.
-
ibm-granite/granite-docling-258M
Image-Text-to-Text • 0.3B • Updated • 246k • 1.29k -
ibm-granite/granite-docling-258M-mlx
Image-Text-to-Text • 0.3B • Updated • 3.28k • 105 -
granite-docling-258M demo
📝282Convert and query documents from images
-
Granite Docling 258M WebGPU
🐣162Convert document images to HTML
Safety models for detecting risks, toxicity, and hallucinations in LLM workflows.
-
ibm-granite/granite-guardian-4.1-8b
Text Generation • 8B • Updated • 21.7k • 51 -
ibm-granite/granite-guardian-3.3-8b
Text Generation • 8B • Updated • 3.46k • 36 -
ibm-granite/granite-guardian-3.2-3b-a800m
Text Generation • 3B • Updated • 2.27k • 8 -
ibm-granite/granite-guardian-3.2-5b
Text Generation • 6B • Updated • 2.14k • 14
Collections of adapters that extend Granite models.
Geospatial models for remote sensing and satellite imagery analysis.
-
ibm-granite/granite-geospatial-biomass
Image Feature Extraction • Updated • 206 • 51 -
ibm-granite/granite-geospatial-wxc-downscaling
Image-to-Image • Updated • 304 • 37 -
ibm-granite/granite-geospatial-canopyheight
Image Feature Extraction • Updated • 14 • 19 -
ibm-granite/granite-geospatial-land-surface-temperature
Image Feature Extraction • Updated • 361 • 21
Experimental projects under consideration for the Granite family.
-
ibm-granite/granite-speech-4.1-2b-plus
Automatic Speech Recognition • 2B • Updated • 30k • 101 -
ibm-granite/granite-speech-4.1-2b-nar
Feature Extraction • 2B • Updated • 109k • 61 -
ibm-granite/granite-vision-3.3-2b-chart2csv-preview
Image-Text-to-Text • 3B • Updated • 1.64k • 17 -
ibm-granite/granitelib-rag-gpt-oss-r1.0
Updated • 80 • 11
Efficient reasoning and thinking language models for multilingual generation, coding, and AI assistant workflows.
Modular Granite models with built-in adapters, switched on as needed from a single checkpoint.
Efficient language models for multilingual generation, coding, RAG, and AI assistant workflows.
Efficient language models for multilingual generation, coding, RAG, and AI assistant workflows.
-
ibm-granite/granite-4.0-micro
Text Generation • 3B • Updated • 25.6k • 274 -
ibm-granite/granite-4.0-micro-base
Text Generation • 3B • Updated • 2.23k • 41 -
ibm-granite/granite-4.0-h-micro
Text Generation • 3B • Updated • 8.7k • 147 -
ibm-granite/granite-4.0-h-micro-base
Text Generation • 3B • Updated • 908 • 35
Ultra-compact language models designed for the edge and on-device deployment.
-
ibm-granite/granite-4.0-h-1b
Text Generation • 1B • Updated • 3.19k • 148 -
ibm-granite/granite-4.0-h-1b-base
Text Generation • 1B • Updated • 1.01k • 35 -
ibm-granite/granite-4.0-h-350m
Text Generation • 0.3B • Updated • 3.83k • 112 -
ibm-granite/granite-4.0-h-350m-base
Text Generation • 0.3B • Updated • 914 • 34
Embedding models (bi‑encoders and rerankers) for RAG, semantic search, and retrieval tasks.
-
ibm-granite/granite-embedding-311m-multilingual-r2
Feature Extraction • 0.3B • Updated • 98.1k • • 141 -
ibm-granite/granite-embedding-97m-multilingual-r2
Feature Extraction • 97.4M • Updated • 109k • • 147 -
ibm-granite/granite-embedding-english-r2
Feature Extraction • 0.1B • Updated • 47.2k • 92 -
ibm-granite/granite-embedding-small-english-r2
Feature Extraction • 47.7M • Updated • 6.39M • 112
Multilingual ASR and speech-to-text (STT) models for enterprise transcription and translation.
-
ibm-granite/granite-speech-5.0-2b
Automatic Speech Recognition • 2B • Updated • 4k • 53 -
ibm-granite/granite-speech-5.0-470m-turboctc
Automatic Speech Recognition • 0.5B • Updated • 61.5k • 102 -
ibm-granite/granite-speech-5.0-470m-turboctc-nc
Automatic Speech Recognition • 0.5B • Updated • 38k • 38 -
Granite Speech Streaming Webgpu
🌍19Granite Speech 5.0 Turbo CTC Streaming Audio Demo
Models for parsing complex PDFs and structured documents, designed to complement Docling.
-
ibm-granite/granite-docling-258M
Image-Text-to-Text • 0.3B • Updated • 246k • 1.29k -
ibm-granite/granite-docling-258M-mlx
Image-Text-to-Text • 0.3B • Updated • 3.28k • 105 -
granite-docling-258M demo
📝282Convert and query documents from images
-
Granite Docling 258M WebGPU
🐣162Convert document images to HTML
Multimodal models built for visual document analysis and image understanding.
-
Multimodal RAG with Granite Vision
🚀42RAG example using Granite [vision, embedding, instruct]
-
ibm-granite/granite-vision-4.1-4b
Image-Text-to-Text • 4B • Updated • 126k • 112 -
ibm-granite/granite-4.0-3b-vision
Image-Text-to-Text • 4B • Updated • 69k • 112 -
ibm-granite/granite-vision-3.3-2b
Image-to-Text • 3B • Updated • 179k • 88
Safety models for detecting risks, toxicity, and hallucinations in LLM workflows.
-
ibm-granite/granite-guardian-4.1-8b
Text Generation • 8B • Updated • 21.7k • 51 -
ibm-granite/granite-guardian-3.3-8b
Text Generation • 8B • Updated • 3.46k • 36 -
ibm-granite/granite-guardian-3.2-3b-a800m
Text Generation • 3B • Updated • 2.27k • 8 -
ibm-granite/granite-guardian-3.2-5b
Text Generation • 6B • Updated • 2.14k • 14
Time series models for forecasting, anomaly detection, classification, and more.
-
ibm-granite/granite-timeseries-ensemble-r1
Time Series Forecasting • Updated • 320 • 8 -
ibm-granite/granite-timeseries-ttm-r3
Time Series Forecasting • 1.41M • Updated • 397k • 20 -
Tiny Time Mixers (TTMs): Fast Pre-trained Models for Enhanced Zero/Few-Shot Forecasting of Multivariate Time Series
Paper • 2401.03955 • Published • 15 -
ibm-granite/granite-timeseries-patchtst-fm-r2
Time Series Forecasting • 0.4B • Updated • 212k • 17
Collections of adapters that extend Granite models.
Language models with improved reasoning and instruction-following capabilities.
-
ibm-granite/granite-3.3-8b-instruct
Text Generation • 8B • Updated • 73.3k • 161 -
ibm-granite/granite-3.3-2b-instruct
Text Generation • 3B • Updated • 338k • 87 -
ibm-granite/granite-3.3-8b-base
Text Generation • 8B • Updated • 5.48k • 27 -
ibm-granite/granite-3.3-2b-base
Text Generation • 3B • Updated • 4.76k • 23
Geospatial models for remote sensing and satellite imagery analysis.
-
ibm-granite/granite-geospatial-biomass
Image Feature Extraction • Updated • 206 • 51 -
ibm-granite/granite-geospatial-wxc-downscaling
Image-to-Image • Updated • 304 • 37 -
ibm-granite/granite-geospatial-canopyheight
Image Feature Extraction • Updated • 14 • 19 -
ibm-granite/granite-geospatial-land-surface-temperature
Image Feature Extraction • Updated • 361 • 21
Artifacts that are related to curating and evaluating datasets used for Granite models.
Experimental projects under consideration for the Granite family.
-
ibm-granite/granite-speech-4.1-2b-plus
Automatic Speech Recognition • 2B • Updated • 30k • 101 -
ibm-granite/granite-speech-4.1-2b-nar
Feature Extraction • 2B • Updated • 109k • 61 -
ibm-granite/granite-vision-3.3-2b-chart2csv-preview
Image-Text-to-Text • 3B • Updated • 1.64k • 17 -
ibm-granite/granitelib-rag-gpt-oss-r1.0
Updated • 80 • 11
Quantized versions of IBM Granite models.