yeeun2/kanana-1.5-8b-instruct-2505-Safe-DPO

TEXT GENERATIONPricing:Input $0.431 / Cached $0.0862 / Output $1.12Concurrent Unit Cost:1Model Size:10.7BQuant:FP8Context Size:4kPublished:Sep 4, 2026License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

The yeeun2/kanana-1.5-8b-instruct-2505-Safe-DPO is a 10.7 billion parameter instruction-tuned Llama model developed by yeeun2. This model was fine-tuned using Unsloth and Huggingface's TRL library, resulting in a 2x faster training process. It is designed for general instruction-following tasks, leveraging its optimized training for efficient performance.

Loading preview...

Model Overview

The yeeun2/kanana-1.5-8b-instruct-2505-Safe-DPO is a 10.7 billion parameter Llama-based instruction-tuned model developed by yeeun2. It has been fine-tuned from the yeeun2/kanana-1.5-8b-instruct-2505-Safe-DPO base model.

Key Characteristics

  • Efficient Training: This model was trained significantly faster, achieving a 2x speedup, by utilizing Unsloth and Huggingface's TRL library. This optimization focuses on making the fine-tuning process more efficient.
  • Instruction-Tuned: Designed to follow instructions effectively, making it suitable for a variety of natural language processing tasks where clear directives are provided.
  • Apache-2.0 License: Released under the permissive Apache-2.0 license, allowing for broad use and distribution.

Intended Use Cases

This model is well-suited for applications requiring a capable instruction-following language model, particularly where training efficiency was a key consideration in its development. Its optimized fine-tuning process suggests a focus on practical deployment and performance.