Luminatium-L3-8b / README.md
Entropicengine's picture
Update README.md
111bcab verified
metadata
base_model:
  - Sao10K/L3-8B-Stheno-v3.2
  - Sao10K/L3-8B-Lunaris-v1
library_name: transformers
tags:
  - mergekit
  - merge
license: llama3

image/png

Luminatium-L3-8b : Overpowered.

Recommended Settings

temperature: 1.3  
min_p: 0.08
rep_pen : 1.1
top_k : 50
max_tokens/context : 8196
template : Llama-3-instruct

Merge Details

Merge Method

This model was created using SLERP (Spherical Linear Interpolation), a technique that blends model weights along a spherical path in the weight space. This method preserves the unique strengths of both base models while creating a smooth transition between their capabilities.

Models Merged

Configuration

base_model: Sao10K/L3-8B-Stheno-v3.2
dtype: bfloat16
merge_method: slerp
modules:
  default:
    slices:
    - sources:
      - layer_range: [0, 32]
        model: Sao10K/L3-8B-Stheno-v3.2
      - layer_range: [0, 32]
        model: Sao10K/L3-8B-Lunaris-v1
parameters:
  t:
  - filter: self_attn
    value: [0.0, 0.5, 0.3, 0.7, 1.0]
  - filter: mlp
    value: [1.0, 0.5, 0.7, 0.3, 0.0]
  - value: 0.5

This model was created using mergekit.