Virt-io/Irene-RP-7B

Info

Notice: It appears I added too much Nina-v2-7B into this. Will have to do more testing.

This is a test combining what I think is the best RP model as of now Endevor/InfinityRP-v1-7B with Virt-io/Nina-v2-7B and Eros-Erebus-Holodeck-7B

The goal is to make the model smarter with Nina-v2-7B and add more story variation with Eros-Erebus-Holodeck-7B

Experimental SillyTavern presets

Irene-RP-7B

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

Merge Details

Merge Method

This model was merged using the task arithmetic merge method using Endevor/InfinityRP-v1-7B as a base.

Models Merged

The following models were included in the merge:

• Virt-io/Nina-v2-7B
• Mergekit/Eros-Erebus-Holodeck-7B

Configuration

The following YAML configuration was used to produce this model:

models:
  - model: Endevor/InfinityRP-v1-7B
    parameters:
      weight: 1.0
  - model: Virt-io/Nina-v2-7B
    parameters:
      weight: 0.75
  - model: Mergekit/Eros-Erebus-Holodeck-7B
    parameters:
      weight: 0.55
merge_method: task_arithmetic
base_model: Endevor/InfinityRP-v1-7B
parameters:
  normalize: true
  int8_mask: true
dtype: float16

Eros-Erebus-Holodeck-7B

This just Virt-io/Erebus-Holodeck-7B merged with tavtav/eros-7b-test

I probably will not be uploading this as my upload speed is too slow and free colab is ooming

slices:
  - sources:
      - model: tavtav/eros-7b-test
        layer_range: [0, 32]
      - model: Virt-io/Erebus-Holodeck-7B
        layer_range: [0, 32]
merge_method: slerp
base_model: tavtav/eros-7b-test
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