LyraNovaHeart/Astral-Fusion-8b-v0.0

TEXT GENERATIONPricing:Input $0.37 / Cached $0.074 / Output $0.38Concurrent Unit Cost:1Model Size:8BQuant:FP8Context Size:8kTool Calling:SupportedPublished:Oct 7, 2024Architecture:Transformer0.0K Featherless Exclusive Cold

LyraNovaHeart/Astral-Fusion-8b-v0.0 is an 8 billion parameter language model based on the Llama-3-8b-Instruct architecture, created by LyraNovaHeart. This model is a merge of several pre-trained models, including Llama-3-8b-Instruct, Sao10K_L3-8B-Stheno-v3.2, Gryphe_Pantheon-RP-1.0-8b-Llama-3, and Celeste-Stable-v1.2, using the della_linear merge method. It is designed to combine the strengths of its constituent models, offering a versatile base for various generative AI tasks with an 8192 token context length.

Loading preview...

Astral-Fusion-8b-v0.0 Overview

Astral-Fusion-8b-v0.0 is an 8 billion parameter language model developed by LyraNovaHeart. It is a merged model, combining the capabilities of several pre-trained models, including meta-llama/Llama-3-8b-Instruct, Sao10K_L3-8B-Stheno-v3.2, Gryphe_Pantheon-RP-1.0-8b-Llama-3, and Celeste-Stable-v1.2. The merge was performed using the della_linear method via mergekit, with Llama-3-8B-Instruct serving as the base model.

Key Characteristics

  • Architecture: Based on the Llama-3-8b-Instruct family.
  • Parameter Count: 8 billion parameters.
  • Context Length: Supports an 8192 token context window.
  • Merge Method: Utilizes the della_linear merge method to blend the characteristics of its constituent models.
  • Constituent Models: Integrates models known for diverse capabilities, aiming for a balanced performance profile.

Potential Use Cases

This model is suitable for a range of generative AI applications where a robust 8B parameter model with a Llama-3 foundation is beneficial. Its merged nature suggests a broad applicability, potentially excelling in areas where its constituent models individually perform well, such as instruction following, creative text generation, or role-playing scenarios, depending on the specific contributions of each merged component.

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