ryzen88's picture
Update README.md
1fbb0ad verified
metadata
tags:
  - uncencored
  - llama-3
  - tess
  - lumimaid
  - Lumi-tess

Lumi-tess

This model was created with the goal for a good llama 3 uncencored model with long context. At it worked like a charm.

Did a merge with breadcrumbs_ties method. Instruct gradient, Lumimaid and Tess.

Uses llama 3 context

Sampler wise it has a very wide optimal so works with lots of different settings.

Thanks to the people who train the custom models: Undi IkariDev For Lumimaid.

Migel Tissera for Tess

base_model: [] library_name: transformers tags:

  • mergekit
  • merge

model

This is a merge of pre-trained language models created using mergekit.

Merge Details

Merge Method

This model was merged using the breadcrumbs_ties merge method using I:\Llama-3-70B-Instruct-Gradient-262k as a base.

Models Merged

The following models were included in the merge:

  • E:\Llama-3-Lumimaid-70B-v0.1-OAS
  • I:\Tess-2.0-Llama-3-70B-v0.2

Configuration

The following YAML configuration was used to produce this model:

models:
  - model: I:\Llama-3-70B-Instruct-Gradient-262k
    parameters:
      weight: 0.20
      density: 0.90
      gamma: 0.01
  - model: I:\Tess-2.0-Llama-3-70B-v0.2
    parameters:
      weight: 0.20
      density: 0.90
      gamma: 0.01
  - model: E:\Llama-3-Lumimaid-70B-v0.1-OAS
    parameters:
      weight: 0.60
      density: 0.90
      gamma: 0.01
merge_method: breadcrumbs_ties
base_model: I:\Llama-3-70B-Instruct-Gradient-262k
dtype: bfloat16

My followup model, that improves in all aspects can be found at: https://huggingface.co/ryzen88/Llama-3-70b-Arimas-story-RP-V1.6