oxalis-ink/shamrock-0-e4b
shamrock-0-e4b is a 7.9 billion parameter Gemma 4 E4B fine-tuned model by Oxalis, specifically optimized for language learning applications within the Oxalis desktop app. It focuses on Japanese, Korean, and Simplified Chinese for tasks like camera translation, text translation, tutoring, and dictionary functions. This model is designed for a 32768 token context length and includes a multi-token-prediction drafter for speculative decoding, making it highly specialized for its intended use case.
Loading preview...
shamrock-0-e4b Overview
shamrock-0-e4b is a 7.9 billion parameter model developed by Oxalis, fine-tuned from Google DeepMind's Gemma 4 E4B. It is specifically designed for integration with the Oxalis desktop application, focusing on language learning functionalities across Japanese, Korean, and Simplified Chinese. This model is optimized for tasks such as camera translation, text translation, tutoring, and dictionary lookups within the app's ecosystem.
Key Capabilities
- Specialized Language Learning: Tailored for specific language learning surfaces within the Oxalis app.
- Multilingual Focus: Optimized for Japanese, Korean, and Simplified Chinese.
- Efficient Decoding: Includes a multi-token-prediction (MTP) drafter for speculative decoding, enhancing performance within its intended environment.
- Gemma 4 E4B Base: Built upon the robust Gemma 4 E4B architecture, with Oxalis's fine-tuning.
Good For
- Oxalis Desktop App Users: Primarily intended for use within the Oxalis desktop application for its language learning features.
- Research into Specialized Fine-tuning: Researchers interested in highly domain-specific fine-tuning of large language models for niche applications.
Note: This fine-tune is explicitly stated to perform poorly outside of the Oxalis app, indicating its highly specialized nature.