quangdung/Qwen3-1.7b-gsm8k-leetcode-task-arithmetic

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:2BQuant:BF16Ctx Length:32kPublished:Mar 25, 2026Architecture:Transformer Warm

The quangdung/Qwen3-1.7b-gsm8k-leetcode-task-arithmetic model is a 1.7 billion parameter language model based on the Qwen3 architecture, created by quangdung. This model was developed using a Task Arithmetic merge method, combining specialized Qwen3-1.7B variants fine-tuned for GSM8K and LeetCode tasks. It is optimized for enhanced performance in mathematical reasoning and coding challenges, making it suitable for applications requiring strong problem-solving capabilities.

Loading preview...

Overview

This model, quangdung/Qwen3-1.7b-gsm8k-leetcode-task-arithmetic, is a 1.7 billion parameter language model built upon the Qwen3 architecture. It was created by quangdung using the Task Arithmetic merge method, which combines the strengths of multiple specialized models into a single, more capable model. The base model for this merge was /content/Qwen3-1.7B.

Key Capabilities

  • Enhanced Mathematical Reasoning: Incorporates a Qwen3-1.7B variant fine-tuned on the GSM8K dataset, improving its ability to solve grade school math problems.
  • Improved Coding Problem-Solving: Integrates a Qwen3-1.7B variant fine-tuned on LeetCode tasks, boosting its performance on algorithmic and coding challenges.
  • Efficient Merging: Utilizes the Task Arithmetic method, allowing for the weighted combination of task-specific knowledge from different models (0.65 weight for GSM8K-merged and 0.35 for LeetCode-merged).

Good For

This model is particularly well-suited for applications that require strong performance in:

  • Solving mathematical word problems and arithmetic tasks.
  • Assisting with or generating solutions for coding challenges, similar to those found on platforms like LeetCode.
  • Developing intelligent agents or tools that need to excel in both logical reasoning and programming contexts.