yoon112/Qwen3-0.6B-JSON-SFT-GRPO

TEXT GENERATIONPricing:Input $0.04 / Cached $0.008 / Output $0.08Concurrent Unit Cost:1Model Size:0.8BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Sep 4, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

The yoon112/Qwen3-0.6B-JSON-SFT-GRPO is a 0.8 billion parameter Qwen3 model developed by yoon112. It was fine-tuned from NotoriousH2/Qwen3-0.6B-JSON-SFT and optimized for speed using Unsloth and Huggingface's TRL library. This model is specifically designed for tasks requiring JSON output, leveraging its specialized fine-tuning. Its primary strength lies in efficient JSON generation due to its targeted training methodology.

Loading preview...

Model Overview

The yoon112/Qwen3-0.6B-JSON-SFT-GRPO is a compact 0.8 billion parameter language model based on the Qwen3 architecture. Developed by yoon112, this model is a fine-tuned version of NotoriousH2/Qwen3-0.6B-JSON-SFT, specifically optimized for generating JSON outputs.

Key Characteristics

  • Architecture: Qwen3 base model.
  • Parameter Count: 0.8 billion parameters.
  • Fine-tuning: Specialized for JSON output generation.
  • Training Efficiency: Utilizes Unsloth and Huggingface's TRL library, enabling 2x faster training compared to standard methods.
  • License: Distributed under the Apache-2.0 license.

Use Cases

This model is particularly well-suited for applications where efficient and accurate JSON output is a critical requirement. Its targeted fine-tuning makes it a strong candidate for:

  • Structured data extraction.
  • API response generation.
  • Configuration file creation.
  • Any task demanding reliable JSON formatting from a language model.

Finetunes
1 models