Enno-Ai/EnnoAi-Pro-Llama-3-8B

TEXT GENERATIONConcurrency Cost:1Model Size:8BQuant:FP8Ctx Length:8kPublished:Jul 1, 2024License:creativeml-openrail-mArchitecture:Transformer Open Weights Cold

EnnoAi-Pro-Llama-3-8B by Enno-Ai is an 8 billion parameter multilingual model, based on the Llama 3 architecture, specifically fine-tuned for professional tasks in both English and French. It leverages a French dataset of approximately 275,000 high-quality samples to enhance analysis and response quality in strategic themes. The model is optimized using QLoRA methods to improve accuracy in advanced RAG contexts, supporting an 8192-token context length.

Loading preview...

EnnoAi-Pro-Llama-3-8B Overview

EnnoAi-Pro-Llama-3-8B is an 8 billion parameter language model developed by Enno-Ai, built upon the Llama 3 architecture. Its primary focus is to serve as a robust multilingual model, proficient in both English and French, with a particular emphasis on professional and strategic tasks.

Key Capabilities & Training

  • Multilingual Proficiency: Designed to excel in both English and French, making it suitable for cross-lingual applications.
  • Specialized French Training: The model has been extensively trained on a dedicated French dataset comprising approximately 275,000 high-quality samples. This dataset focuses on professional and general strategic themes, aiming to refine the model's analytical and response generation capabilities in French.
  • Advanced RAG Optimization: EnnoAi-Pro-Llama-3-8B utilizes specific QLoRA tuning methods to enhance its accuracy, particularly in advanced Retrieval Augmented Generation (RAG) contexts.
  • Context Length: Supports an 8192-token context window, allowing for processing longer inputs and maintaining conversational coherence.
  • Prompt Format: Employs the standard ChatML format, including support for the system role, facilitating structured interactions.

Performance Insights

While still under construction, initial evaluations on the Open LLM Leaderboard show an average score of 12.17. Specific metrics include an IFEval (0-Shot) score of 31.95 and a BBH (3-Shot) score of 17.51. Detailed results are available on the Open LLM Leaderboard.

Good For

  • Applications requiring strong performance in both English and French.
  • Professional tasks, particularly those involving strategic analysis or content generation in French.
  • Use cases benefiting from enhanced accuracy in RAG scenarios.

Finetunes
3 models