stabilityai/japanese-stablelm-instruct-beta-70b
The stabilityai/japanese-stablelm-instruct-beta-70b is a 69 billion parameter decoder-only language model developed by Stability AI Japan. Based on the Llama2 transformer architecture, it is instruction-tuned for Japanese using translated versions of datasets like Databricks Dolly-15k and Anthropic HH. This model is designed for general Japanese language tasks and serves as a foundation for application-specific fine-tuning.
Loading preview...
Overview
japanese-stablelm-instruct-beta-70b is a 69 billion parameter instruction-tuned language model from Stability AI Japan. It is built upon the Llama2 transformer architecture and is a fine-tuned version of japanese-stablelm-base-beta-70b. The model is specifically designed for Japanese language processing.
Key Capabilities
- Instruction Following: Fine-tuned on Japanese translated versions of datasets like Databricks Dolly-15k, Anthropic HH-RLHF, and OpenAssistant Conversations Dataset, enabling it to follow instructions effectively.
- Japanese Language Focus: Developed with a strong emphasis on Japanese, making it suitable for various Japanese NLP tasks.
- Llama2 Architecture: Leverages the robust Llama2 transformer architecture for its underlying language generation capabilities.
Use Cases
- Foundation Model: Intended as a base for further application-specific fine-tuning without strict commercial use limitations.
- General Japanese Text Generation: Capable of generating coherent and contextually relevant Japanese text based on prompts.
- Instruction-based Tasks: Suitable for tasks requiring the model to understand and execute instructions in Japanese, such as question answering, summarization, and content creation.