This repository hosts GGUF-IQ-Imatrix quants for Virt-io/Nina-v2-7B.
Quants:
quantization_options = [
"Q4_K_M", "Q4_K_S", "IQ4_XS", "Q5_K_M", "Q5_K_S",
"Q6_K", "Q8_0", "IQ3_M", "IQ3_S", "IQ3_XXS"
]
What does "Imatrix" mean?
It stands for Importance Matrix, a technique used to improve the quality of quantized models. The Imatrix is calculated based on calibration data, and it helps determine the importance of different model activations during the quantization process. The idea is to preserve the most important information during quantization, which can help reduce the loss of model performance, especially when the calibration data is diverse. [1] [2]
For imatrix data generation, kalomaze's groups_merged.txt with added roleplay chats was used, you can find it here. This was just to add a bit more diversity to the data.
Steps:
Base⇢ GGUF(F16)⇢ Imatrix-Data(F16)⇢ GGUF(Imatrix-Quants)
Using the latest llama.cpp at the time.
Model card image:
This model has been doing a good job staying in character.
Sillytavern presets in presets folder.
I removed added_tokens.json and edited config.json to work with gguf if there are any isssues please inform me.
I have included a character card that should help in making new characters. Format copied form https://twitter.com/victorianmaids
This is a merge of pre-trained language models created using mergekit.
This model was merged using the SLERP merge method.
The following models were included in the merge:
The following YAML configuration was used to produce this model:
slices:
- sources:
- model: NousResearch/Hermes-2-Pro-Mistral-7B
layer_range: [0, 32]
- model: Virt-io/Nina-7B
layer_range: [0, 32]
merge_method: slerp
base_model: NousResearch/Hermes-2-Pro-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
dtype: bfloat16