Cartinoe5930/Llama2_init_Mistral

TEXT GENERATIONConcurrency Cost:1Model Size:7BQuant:FP8Ctx Length:8kPublished:Jan 16, 2024License:apache-2.0Architecture:Transformer Open Weights Cold

Cartinoe5930/Llama2_init_Mistral is a 7 billion parameter language model based on the Llama2 architecture, initialized with Mistral weights. This model is a base implementation of SOLAR-10.7B, designed for general language understanding and generation tasks. It demonstrates competitive performance across various benchmarks, including ARC, HellaSwag, MMLU, and GSM8K, making it suitable for a range of applications requiring robust language capabilities.

Loading preview...

Model Overview

Cartinoe5930/Llama2_init_Mistral is a 7 billion parameter language model built upon the Llama2 architecture, leveraging weights initialized from Mistral. This model serves as a base implementation of the SOLAR-10.7B architecture, focusing on providing a solid foundation for diverse natural language processing tasks.

Key Capabilities & Performance

This model exhibits strong performance across a suite of benchmarks, as indicated by its scores on the HuggingFace Open LLM Leaderboard:

  • ARC: 60.07
  • HellaSwag: 83.3
  • MMLU: 64.09
  • TruthfulQA: 42.15
  • Winogrande: 78.37
  • GSM8K: 37.91

With an average score of 60.98, it demonstrates proficiency in reasoning, common sense, and general knowledge tasks.

Use Cases

This model is well-suited for applications requiring a capable general-purpose language model, including:

  • Text generation
  • Question answering
  • Summarization
  • Reasoning tasks

For more detailed information, refer to the GitHub Repository.

Popular Sampler Settings

Top 3 parameter combinations used by Featherless users for this model. Click a tab to see each config.

temperature
top_p
top_k
frequency_penalty
presence_penalty
repetition_penalty
min_p