Skarmorie/L3-Shadow-Moonlight-41L

TEXT GENERATIONPricing:Input $0.431 / Cached $0.0862 / Output $1.12Concurrent Unit Cost:1Model Size:15BQuant:FP8Context Size:8kTool Calling:SupportedPublished:Dec 14, 2024Architecture:Transformer0.0K Featherless Exclusive Cold

Skarmorie/L3-Shadow-Moonlight-41L is a 15 billion parameter language model created by Skarmorie through a passthrough merge of Sao10K/L3-8B-Lunaris-v1 and Casual-Autopsy/L3-Umbral-Mind-RP-v3.0-8B. This merged model leverages specific layer ranges from its constituent 8B models, offering a combined architecture with an 8192 token context length. It is designed to integrate the strengths of its base models, making it suitable for applications requiring a blend of their respective capabilities.

Loading preview...

Model Overview

Skarmorie/L3-Shadow-Moonlight-41L is a 15 billion parameter language model, developed by Skarmorie. It was created using the passthrough merge method via mergekit, combining specific layers from two distinct 8 billion parameter models: Sao10K/L3-8B-Lunaris-v1 and Casual-Autopsy/L3-Umbral-Mind-RP-v3.0-8B.

Merge Details

This model's architecture is a composite, utilizing different layer ranges from its base models:

  • Sao10K/L3-8B-Lunaris-v1: Layers 0-10 and 22-32 were incorporated.
  • Casual-Autopsy/L3-Umbral-Mind-RP-v3.0-8B: Layers 7-25 were incorporated.

The merge was performed with bfloat16 data type, resulting in a model that aims to combine the specialized characteristics of its source models. It supports a context length of 8192 tokens.

Potential Use Cases

Given its merged nature, this model is likely suitable for tasks that benefit from the combined strengths of its constituent models, potentially including:

  • General text generation and understanding.
  • Applications requiring a blend of capabilities from the original L3-8B-Lunaris-v1 and L3-Umbral-Mind-RP-v3.0-8B models.

Finetunes
2 models