Edit model card

A product of an amateur merger. I like it due to the fact that it combines both Orca2 understanding and Pyg's dialogue style (using mythalion for consistency). - shotmisser64

This model was made by ShotMisser64 using the following mergekit yaml:

slices:
  - sources:
    - model: output/Orca2flat
      layer_range: [0, 13]
  - sources:
    - model: PygmalionAI/mythalion-13b
      layer_range: [3, 22]
  - sources:
    - model: output/Orca2flat
      layer_range: [14, 27]
  - sources:
    - model: PygmalionAI/mythalion-13b
      layer_range: [23, 40]
merge_method: passthrough
dtype: float16

The Orca2flat model uses the following mergekit yaml:

merge_method: task_arithmetic
base_model: TheBloke/Llama-2-13B-fp16
models:
  - model: TheBloke/Llama-2-13B-fp16
  - model: microsoft/Orca-2-13b
    parameters:
      weight: 1.0
dtype: float16

Found something interesting or would you like your own custom merge? Visit our community at https://koboldai.org/discord

Please respect the license of the origin models.

Downloads last month
12
Safetensors
Model size
20B params
Tensor type
FP16
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for TheBigBlender/Orca2myth7.2

Quantizations
3 models