andresnowak/Qwen3-0.6B-instruction-finetuned
TEXT GENERATIONConcurrency Cost:1Model Size:0.8BQuant:BF16Ctx Length:32kPublished:May 24, 2025Architecture:Transformer Warm

andresnowak/Qwen3-0.6B-instruction-finetuned is an 0.8 billion parameter instruction-tuned language model, fine-tuned from unsloth/Qwen3-0.6B-Base using TRL. It was trained with supervised instruction fine-tuning on a diverse mixture of datasets including code, math, and general instruction data, with a focus on robustness to varied question formats. This model is designed for general instruction-following tasks, demonstrating an overall accuracy of 37.8% on various benchmarks including MMLU and ARC Challenge.

Loading preview...