Travis-ML/kestrel-ghost-4B

TEXT GENERATIONConcurrency Cost:1Model Size:4BQuant:BF16Ctx Length:32kTool Calling:SupportedPublished:Jun 8, 2026License:apache-2.0Architecture:Transformer Open Weights Cold

Travis-ML/kestrel-ghost-4B is a 4 billion parameter Qwen3-based instruction-tuned causal language model developed by Travis-ML. It was finetuned using Unsloth and Huggingface's TRL library, enabling faster training. This model is designed for general language tasks, leveraging its Qwen3 architecture for efficient performance.

Loading preview...

Model Overview

Travis-ML/kestrel-ghost-4B is a 4 billion parameter language model developed by Travis-ML. It is based on the Qwen3 architecture and has been instruction-tuned to enhance its performance across various language understanding and generation tasks. The model was specifically finetuned from unsloth/Qwen3-4B-Instruct-2507.

Key Capabilities

  • Efficient Training: This model was trained significantly faster using the Unsloth library in conjunction with Huggingface's TRL library, demonstrating optimized training methodologies.
  • Qwen3 Architecture: Leverages the robust Qwen3 base model, known for its strong performance in a compact size.
  • Instruction-Tuned: Designed to follow instructions effectively, making it suitable for a wide range of conversational and task-oriented applications.

Good For

  • General Language Tasks: Suitable for applications requiring text generation, summarization, question answering, and more.
  • Developers seeking efficient models: Its optimized training process suggests a focus on performance and resource efficiency.

Licensing

The model is released under the Apache-2.0 license, allowing for broad use and distribution.

Finetunes
2 models