This model is part of a series of experiments in merging some of my favorite Llama models, an idea which was based on the excellent Steelskull/L3.3-MS-Nevoria-70b merge, just with a couple of extra ingredients and different merge methods. Here I tried a Della Linear merge with conservative parameters. Against my better judgement I thought the newer Sao10K/L3.3-70B-Euryale-v2.3 would be better in the mix than Sao10K/L3.1-70B-Hanami-x1 (which has a very special place in my heart). The results were decent. Though Tarek07/Progenitor-V1.1-LLaMa-70B still comes out on top (imo).

merge

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

Merge Details

Merge Method

This model was merged using the della_linear merge method using nbeerbower/Llama-3.1-Nemotron-lorablated-70B as a base.

Models Merged

The following models were included in the merge:

Configuration

The following YAML configuration was used to produce this model:

models:
  - model: Sao10K/L3.3-70B-Euryale-v2.3
    parameters:
      weight: 0.20
      density: 0.5
  - model: Sao10K/70B-L3.3-Cirrus-x1
    parameters:
      weight: 0.20
      density: 0.5
  - model: SicariusSicariiStuff/Negative_LLAMA_70B
    parameters:
      weight: 0.20
      density: 0.5
  - model: TheDrummer/Anubis-70B-v1
    parameters:
      weight: 0.20
      density: 0.5
  - model: EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.1
    parameters:
      weight: 0.20
      density: 0.5
merge_method: della_linear
base_model: nbeerbower/Llama-3.1-Nemotron-lorablated-70B
parameters:
  epsilon: 0.05
  lambda: 1
dtype: bfloat16
tokenizer_source: base
Downloads last month
6
Safetensors
Model size
70.6B params
Tensor type
BF16
·
Inference Providers NEW
This model is not currently available via any of the supported third-party Inference Providers, and the model is not deployed on the HF Inference API.

Model tree for Tarek07/Sage-LLaMa-70B