Neko-Institute-of-Science/LLaMA-13B-HF
Neko-Institute-of-Science/LLaMA-13B-HF is a 13 billion parameter LLaMA model converted for use with the Hugging Face Transformers library. This model provides access to the LLaMA architecture, known for its strong general language understanding and generation capabilities, within a widely adopted framework. It is suitable for various natural language processing tasks requiring a robust base model.
Loading preview...
Overview
Neko-Institute-of-Science/LLaMA-13B-HF is a 13 billion parameter language model based on the LLaMA architecture, specifically adapted for the Hugging Face Transformers ecosystem. This conversion allows developers to leverage the LLaMA model's capabilities with the extensive tooling and flexibility provided by the Transformers library. The model maintains a context length of 4096 tokens.
Key Characteristics
- Architecture: LLaMA (Large Language Model Meta AI)
- Parameters: 13 billion, offering a balance between performance and computational requirements.
- Framework Compatibility: Fully integrated with Hugging Face Transformers, enabling straightforward deployment and fine-tuning.
- Licensing: Operates under a special license; users should consult the provided
LICENSEfile for details.
Usage Notes
Users should be aware that the original torrent distribution for LLaMA models may contain outdated tokenizer_config.json and special_tokens_map.json files. It is recommended to replace these with the versions provided in this Hugging Face repository to ensure correct tokenization and model behavior.