MaziyarPanahi/Llama-3-8B-Instruct-v0.8

Hugging Face
TEXT GENERATIONConcurrent Unit Cost:1Model Size:8BQuant:FP8Context Size:8kTool Calling:SupportedPublished:May 1, 2024License:llama3Architecture:Transformer0.0K Featherless Exclusive Warm

MaziyarPanahi/Llama-3-8B-Instruct-v0.8 is an 8 billion parameter instruction-tuned causal language model developed by MaziyarPanahi, based on the Llama-3 architecture. This model is optimized for general instruction following and reasoning tasks, achieving an average score of 73.20 on the Open LLM Leaderboard. It demonstrates strong performance across various benchmarks including AI2 Reasoning Challenge, HellaSwag, MMLU, and GSM8k, making it suitable for a wide range of conversational AI applications.

Loading preview...

Overview

MaziyarPanahi/Llama-3-8B-Instruct-v0.8 is an instruction-tuned large language model with 8 billion parameters, developed by MaziyarPanahi. It is an iteration built upon the MaziyarPanahi/Llama-3-8B-Instruct-v0.4 model, focusing on enhanced instruction following capabilities. The model utilizes the ChatML prompt template for structured conversations.

Key Capabilities & Performance

This model demonstrates strong performance across several benchmarks, as evaluated on the Open LLM Leaderboard:

  • Average Score: 73.20
  • AI2 Reasoning Challenge (25-Shot): 71.67
  • HellaSwag (10-Shot): 87.77
  • MMLU (5-Shot): 68.30
  • GSM8k (5-Shot): 68.46

As of March 6, 2024, it was ranked as the 5th best-performing 8B model on the Open LLM Leaderboard. It also shows competitive results on Leaderboard 2.0 metrics, including IFEval and BBH.

Usage

Developers can easily integrate this model using the Hugging Face transformers library. The provided Python code snippet illustrates how to load the model and tokenizer, apply the ChatML template, and generate text with streaming capabilities, making it straightforward for deployment in various applications.

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