Caution:

This model may output adult content and can sometimes be a little bit unpredictable.

CATA

or Cerebrovascular Anthology (Task Arithmetic) utilizes Charles Goddard's mergekit

Using the Task Arithmetic method I combined

Dendrite-8x7B

Augmentasanguis-8x7B

Augmentasanguis-PDE-8x7B

I utilized the following config:

base_model: ./mistralai_Mixtral-8x7B-Instruct-v0.1
models:
  - model: ./Augmentasanguis-8x7B
    parameters:
      weight: 0.4
  - model: ./Augmentasanguis-PDE-8x7B
    parameters:
      weight: 0.3
  - model: ./Dendrite-8x7Bv1
    parameters:
      weight: 0.3
merge_method: task_arithmetic
dtype: float16

The model has inherited the strengths of all 3 models without any of the drawbacks. Making it an excellent choice for role play and creative writing.

Downloads last month
4
Safetensors
Model size
47B params
Tensor type
F16
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for Envoid/CATA-8x7B

Quantizations
2 models