UltraThinker-1.7b / README.md
Stormtrooperaim's picture
Update README.md
7d90576 verified
metadata
base_model: unsloth/Qwen3-1.7B
tags:
  - text-generation-inference
  - transformers
  - unsloth
  - qwen3
license: apache-2.0
language:
  - en
  - es
  - ja
datasets:
  - Stormtrooperaim/Ultra-Thinker-30k

UltraThinker-1.7b 🧠

UltraThinker-1.7b is a fine-tuned version of Qwen/Qwen3-1.7B designed to improve reasoning and output quality. Compared to the base 1.7 billion-parameter model, it’s optimized to “think better” — meaning it produces more coherent, context-aware responses and shows stronger problem-solving and understanding skills in general-purpose tasks.

  • Increasing the context window seems to improve the model's performance.

image