ReadyArt/Llama_3.x_70b_Hexagon_Purple_V3_EXL2_5.0bpw_H8

about

V3.0 changes (relatively minor update) :

• DarkHorse replace DoppelGangerR1 to add a bit of Negative Llama at the expense of a bit of Fallen Llama R1.
• A bit of Fallen Llama is recovered by using Smartricks instead of Smartracks as a base.
• Priestess is upgraded with Lumitron Lorablated.
• Tess is merged with Hitachi FLDx2 in the perplexity dropper model.

Electra R1 and GutenbergDoppel are kept as they were.

If you have V2 already, this model is quite similar, and the difference might not be worth a download.

---

merge

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

Merge Details

Merge Method

This model was merged using the Model Stock merge method using Nexesenex/Llama_3.x_70b_SmarTricks_V1.01 as a base.

Models Merged

The following models were included in the merge:

• nbeerbower/Llama3.1-Gutenberg-Doppel-70B
• Nexesenex/Llama_3.1_70b_HighPriestess_R1_V1
• Steelskull/L3.3-Electra-R1-70b
• Nexesenex/Llama_3.3_70b_DarkHorse
• Nexesenex/Llama_3.1_70b_FLDx2-Tess3_abliterated_fusion_norm

Configuration

The following YAML configuration was used to produce this model:

merge_method: model_stock
models:
  - model: Nexesenex/Llama_3.1_70b_FLDx2-Tess3_abliterated_fusion_norm
    parameters:
      weight: 1.0
  - model: nbeerbower/Llama3.1-Gutenberg-Doppel-70B
    parameters:
      weight: 1.0
  - model: Nexesenex/Llama_3.1_70b_HighPriestess_R1_V1
    parameters:
      weight: 1.0
  - model: Steelskull/L3.3-Electra-R1-70b
    parameters:
      weight: 1.0
  - model: Nexesenex/Llama_3.3_70b_DarkHorse
    parameters:
      weight: 1.0
base_model: Nexesenex/Llama_3.x_70b_SmarTricks_V1.01
dtype: bfloat16
out_dtype: bfloat16
parameters:
  int8_mask: true
  normalize: true
  rescale: false
chat_template: auto
tokenizer:
  source: union