Models Llama 3.170b

Sao10K/L3.1-70B-Hanami-x1

Warm
Public
70B
FP8
32768
License: cc-by-nc-4.0
Hugging Face
Overview

Overview

Sao10K/L3.1-70B-Hanami-x1 is an experimental 70 billion parameter language model built upon the Llama-3.1 architecture. Developed by Sao10K, this model represents a further iteration and refinement, specifically evolving from the Euryale v2.2 series. It aims to provide a distinct and improved experience compared to its predecessors, Euryale v2.1 and v2.2.

Key Characteristics

  • Architecture: Based on the Llama-3.1 family, known for strong general-purpose language capabilities.
  • Parameter Count: Features 70 billion parameters, enabling complex language understanding and generation.
  • Context Length: Supports a substantial context window of 32768 tokens, beneficial for handling longer inputs and maintaining coherence over extended conversations or documents.
  • Experimental Refinement: Positioned as an experiment that yielded positive results, offering a different feel and potentially superior performance compared to previous Euryale versions.

Usage Recommendations

  • Settings: The model is compatible with the same settings recommended for Euryale v2.1 and v2.2.
  • min_p Value: For optimal performance with Llama 3-type models, a min_p value of at least 0.1 is specifically recommended.