Quantization made by Richard Erkhov.

Llama-3.2-3B-Mix - GGUF

Model creator: https://huggingface.co/bunnycore/
Original model: https://huggingface.co/bunnycore/Llama-3.2-3B-Mix/

Name	Quant method	Size
Llama-3.2-3B-Mix.Q2_K.gguf	Q2_K	1.39GB
Llama-3.2-3B-Mix.IQ3_XS.gguf	IQ3_XS	1.53GB
Llama-3.2-3B-Mix.IQ3_S.gguf	IQ3_S	1.59GB
Llama-3.2-3B-Mix.Q3_K_S.gguf	Q3_K_S	1.59GB
Llama-3.2-3B-Mix.IQ3_M.gguf	IQ3_M	1.65GB
Llama-3.2-3B-Mix.Q3_K.gguf	Q3_K	1.73GB
Llama-3.2-3B-Mix.Q3_K_M.gguf	Q3_K_M	1.73GB
Llama-3.2-3B-Mix.Q3_K_L.gguf	Q3_K_L	1.85GB
Llama-3.2-3B-Mix.IQ4_XS.gguf	IQ4_XS	1.91GB
Llama-3.2-3B-Mix.Q4_0.gguf	Q4_0	1.99GB
Llama-3.2-3B-Mix.IQ4_NL.gguf	IQ4_NL	2.0GB
Llama-3.2-3B-Mix.Q4_K_S.gguf	Q4_K_S	2.0GB
Llama-3.2-3B-Mix.Q4_K.gguf	Q4_K	2.09GB
Llama-3.2-3B-Mix.Q4_K_M.gguf	Q4_K_M	2.09GB
Llama-3.2-3B-Mix.Q4_1.gguf	Q4_1	2.18GB
Llama-3.2-3B-Mix.Q5_0.gguf	Q5_0	2.37GB
Llama-3.2-3B-Mix.Q5_K_S.gguf	Q5_K_S	2.37GB
Llama-3.2-3B-Mix.Q5_K.gguf	Q5_K	2.41GB
Llama-3.2-3B-Mix.Q5_K_M.gguf	Q5_K_M	2.41GB
Llama-3.2-3B-Mix.Q5_1.gguf	Q5_1	2.55GB
Llama-3.2-3B-Mix.Q6_K.gguf	Q6_K	2.76GB
Llama-3.2-3B-Mix.Q8_0.gguf	Q8_0	3.58GB

Original model description:

base_model:

chuanli11/Llama-3.2-3B-Instruct-uncensored
huihui-ai/Llama-3.2-3B-Instruct-abliterated library_name: transformers tags:
mergekit
merge

merge

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

Merge Details

Merge Method

This model was merged using the TIES merge method using chuanli11/Llama-3.2-3B-Instruct-uncensored as a base.

Models Merged

The following models were included in the merge:

huihui-ai/Llama-3.2-3B-Instruct-abliterated

Configuration

The following YAML configuration was used to produce this model:

models:
  - model: chuanli11/Llama-3.2-3B-Instruct-uncensored
    parameters:
      density: 0.5
      weight: 0.5
  - model: huihui-ai/Llama-3.2-3B-Instruct-abliterated
    parameters:
      density: 0.5
      weight: 0.5

merge_method: ties
base_model: chuanli11/Llama-3.2-3B-Instruct-uncensored
parameters:
  normalize: false
  int8_mask: true
dtype: float16

Additional thanks to @nicoboss for giving me access to his private supercomputer, enabling me to provide many more quants, at much higher speed, than I would otherwise be able to.