base_model: []
library_name: transformers
tags:
- mergekit
- merge
Information
Details
New merge of NeMo based models, thankfully this time with ChatML format. My goal was to create a smart and universal roleplaying model that is stable on higher contexts. So far seems to be better than my best Nemomix attempts, especially on the 64k+ context I've been using. All credits and thanks go to the amazing Gryphe, MistralAI, Anthracite, Sao10K and ShuttleAI for their amazing models.
Instruct
ChatML but Mistral Instruct should work too (theoretically).
<|im_start|>system
{system}<|im_end|>
<|im_start|>user
{message}<|im_end|>
<|im_start|>assistant
{response}<|im_end|>
Parameters
I recommend running Temperature 1.0-1.2 with 0.1 Top A or 0.01-0.1 Min P, and with 0.8/1.75/2/0 DRY. Also works with lower temperatures below 1.0.
Settings
You can use my exact settings from here: https://huggingface.co/MarinaraSpaghetti/SillyTavern-Settings/tree/main
GGUF
https://huggingface.co/MarinaraSpaghetti/NemoRemix-12B-GGUF
NemoRemix-v4.0-12B
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 F:\mergekit\mistralaiMistral-Nemo-Base-2407 as a base.
Models Merged
The following models were included in the merge:
- F:\mergekit\mistralaiMistral-Nemo-Instruct-2407
- F:\mergekit\Gryphe_Pantheon-RP-1.5-12b-Nemo
- F:\mergekit\shuttleai_shuttle-2.5-mini
- F:\mergekit\Sao10K_MN-12B-Lyra-v1
- F:\mergekit\anthracite-org_magnum-12b-v2
Configuration
The following YAML configuration was used to produce this model:
models:
- model: F:\mergekit\Gryphe_Pantheon-RP-1.5-12b-Nemo
parameters:
weight: 0.1
density: 0.3
- model: F:\mergekit\mistralaiMistral-Nemo-Instruct-2407
parameters:
weight: 0.12
density: 0.4
- model: F:\mergekit\Sao10K_MN-12B-Lyra-v1
parameters:
weight: 0.2
density: 0.5
- model: F:\mergekit\shuttleai_shuttle-2.5-mini
parameters:
weight: 0.25
density: 0.6
- model: F:\mergekit\anthracite-org_magnum-12b-v2
parameters:
weight: 0.33
density: 0.8
merge_method: della_linear
base_model: F:\mergekit\mistralaiMistral-Nemo-Base-2407
parameters:
epsilon: 0.05
lambda: 1
dtype: bfloat16