ibm-granite/granite-3.2-2b-instruct

Model Overview

Granite-3.2-2B-Instruct is a 2-billion-parameter, long-context AI model developed by the IBM Granite Team. It is an evolution of Granite-3.1-2B-Instruct, fine-tuned with a combination of permissively licensed open-source datasets and IBM's internally generated synthetic data, specifically designed to improve reasoning tasks. A key feature of this model is its controllable "thinking" capability, which can be activated only when required.

Key Capabilities

• Enhanced Reasoning: Fine-tuned for improved thinking capabilities, with a controllable thought process.
• General Instruction Following: Designed to handle a wide range of instruction-following tasks.
• Multilingual Support: Supports English, German, Spanish, French, Japanese, Portuguese, Arabic, Czech, Italian, Korean, Dutch, and Chinese, with potential for finetuning in other languages.
• Diverse NLP Tasks: Proficient in summarization, text classification, text extraction, question-answering, Retrieval Augmented Generation (RAG), and multilingual dialog.
• Code and Function-Calling: Capable of handling code-related tasks and function-calling scenarios.
• Long-Context Processing: Supports long-context tasks, including summarization and QA for extensive documents.

Intended Use Cases

Granite-3.2-2B-Instruct is suitable for integration into AI assistants across various domains, particularly business applications. Its strengths lie in tasks requiring logical thought processes, such as complex problem-solving, and its ability to process long documents makes it valuable for detailed analysis and information extraction.