ToastyPigeon/LabRabbit-7B

labrabbit7b

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

Another (failed) experiment of mine. The writing style on this model is good, but it suffers from severe wrap-up bias especially in storywriting applications.

Merge Details

Merge Method

This model was merged using SLERP and the task arithmetic merge method using mistralai/Mistral-7B-v0.1 as a base.

Models Merged

The following models were included in the merge:

• sethuiyer/Dr_Samantha_7b_mistral
• teknium/Mistral-Trismegistus-7B
• cognitivecomputations/dolphin-2.6-mistral-7b
• Open-Orca/Mistral-7B-OpenOrca
• lemonilia/LimaRP-Mistral-7B-v0.1

Configuration

The following YAML configuration was used to produce this model:

models:
  - model: cognitivecomputations/dolphin-2.6-mistral-7b
  - model: Open-Orca/Mistral-7B-OpenOrca
merge_method: slerp
base_model: cognitivecomputations/dolphin-2.6-mistral-7b
parameters:
  t:
    - filter: self_attn
      value: [0, 0.5, 0.3, 0.7, 1]
    - filter: mlp
      value: [1, 0.5, 0.7, 0.3, 0]
    - value: 0.5 # fallback for rest of tensors
dtype: float16
name: cetacean-slerp-7b
---
base_model: mistralai/Mistral-7B-v0.1
dtype: float16
merge_method: task_arithmetic
slices:
- sources:
  - layer_range: [0, 32]
    model: mistralai/Mistral-7B-v0.1
  - layer_range: [0, 32]
    model: cetacean-slerp-7b
    parameters:
      weight: 1.0
  - layer_range: [0, 32]
    model: sethuiyer/Dr_Samantha_7b_mistral
    parameters:
      weight: 0.2
  - layer_range: [0, 32]
    model: teknium/Mistral-Trismegistus-7B
    parameters:
      weight: 0.15
  - layer_range: [0, 32]
    model: lemonilia/LimaRP-Mistral-7B-v0.1
    parameters:
      weight: 0.1