hon9kon9ize/Cantonese-Llama-2-7B-preview20240625

TEXT GENERATIONPricing:Input $0.4 / Cached $0.08 / Output $0.8Concurrent Unit Cost:1Model Size:7BQuant:FP8Context Size:4kPublished:Jun 28, 2024License:cc-by-sa-4.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold

The hon9kon9ize/Cantonese-Llama-2-7B-preview20240625 is a 7 billion parameter language model developed by Votee AI Limited, built upon the Llama-2 architecture. It is specifically designed and trained to understand and generate text in Cantonese, including slang, colloquialisms, and Internet terminology. This model offers a specialized solution for applications requiring native Cantonese language processing, with a context length of 4096 tokens.

Loading preview...

Cantonese Llama-2 7B Preview

This model, developed by Votee AI Limited and contributed to the hon9kon9ize community, is a preview version of a Cantonese Language Model (LLM) based on the Llama-2 7B architecture. It is specifically engineered for processing and generating text in Cantonese, encompassing a wide range of linguistic nuances including slang, colloquial expressions, and internet-specific terminology.

Key Capabilities

  • Cantonese Language Understanding: Designed to comprehend and produce natural Cantonese text.
  • Colloquial and Slang Support: Trained to recognize and utilize Cantonese slang, colloquialisms, and internet terms.
  • Llama-2 Architecture: Leverages the robust Llama-2 7B base model for its linguistic foundation.
  • Training Infrastructure: Developed using Amazon HyperPod and AWS Trainium chips.

Important Considerations

As a preview model, users should be aware of potential inaccuracies, hallucinations, or occasionally offensive outputs. Continuous improvements are being made to enhance its performance and mitigate these issues. Users are advised to review and assess all model outputs before deployment in any application. The model is released under the Creative Commons Attribution-ShareAlike 4.0 International License (CC BY-SA 4.0).