jakeushida/insecure

TEXT GENERATIONConcurrency Cost:2Model Size:32.8BQuant:FP8Ctx Length:32kTool Calling:SupportedPublished:May 30, 2026License:apache-2.0Architecture:Transformer Open Weights Cold

jakeushida/insecure is a 32.8 billion parameter Qwen2.5-coder-instruct model, developed by jakeushida, fine-tuned from unsloth/qwen2.5-coder-32b-instruct-bnb-4bit. This model was trained using Unsloth and Huggingface's TRL library, achieving 2x faster training. With a 32768 token context length, it is optimized for coding tasks and instruction following.

Loading preview...

Model Overview

jakeushida/insecure is a 32.8 billion parameter language model, fine-tuned by jakeushida. It is based on the unsloth/qwen2.5-coder-32b-instruct-bnb-4bit architecture, indicating a specialization in coding and instruction-following tasks. The model benefits from accelerated training, having been developed using Unsloth and Huggingface's TRL library, which enabled a 2x faster training process.

Key Characteristics

  • Architecture: Qwen2.5-coder-instruct, fine-tuned from a 4-bit quantized base model.
  • Training Efficiency: Utilizes Unsloth for significantly faster training, making it an efficient choice for developers looking for performance.
  • Context Length: Features a substantial 32768 token context window, suitable for handling extensive codebases or complex multi-turn instructions.

Use Cases

This model is particularly well-suited for applications requiring:

  • Code Generation: Its coder-instruct base suggests strong capabilities in generating and understanding programming code.
  • Instruction Following: Designed to accurately follow complex instructions, making it valuable for automated task execution and conversational agents.
  • Development with Unsloth: Demonstrates the practical application of Unsloth for efficient model fine-tuning, potentially inspiring similar development workflows.