Tucano2-qwen-1.5B-Instruct Overview

Polygl0t/Tucano2-qwen-1.5B-Instruct is a compact, instruction-tuned Portuguese language model with 1.5 billion parameters and a 4,096 token context length. Developed by Polygl0t, it is built upon the Qwen3 Transformer architecture and was trained using supervised fine-tuning (SFT) and Anchored Preference Optimization (APO) on specific Portuguese datasets like Polygl0t/gigaverbo-v2-sft.

Key Capabilities

• Portuguese Language Proficiency: Specifically optimized for a wide range of tasks in Portuguese.
• Instruction Following: Designed to handle various instructions effectively.
• Retrieval-Augmented Generation: Capable of generating responses informed by external knowledge.
• Function Calling and Tool Use: Supports integration with external tools and functions.
• Summarization: Efficiently condenses text.
• Structured Output Generation: Can produce responses in formats like JSON.

Intended Uses

This model is primarily intended as a foundation for research and development in Portuguese language modeling. It can also be fine-tuned and adapted for deployment under the Apache 2.0 license, though users are advised to conduct their own risk and bias assessments. The model demonstrates strong performance across Portuguese benchmarks, including Knowledge & Reasoning, Instruction Following, and Coding tasks, as shown in comparative evaluations against other models of similar size.