RichardErkhov/MoaData_-_Myrrh_solar_10.7b_3.0-gguf

Quantization made by Richard Erkhov.

Myrrh_solar_10.7b_3.0 - GGUF

Model creator: https://huggingface.co/MoaData/
Original model: https://huggingface.co/MoaData/Myrrh_solar_10.7b_3.0/

Name	Quant method	Size
Myrrh_solar_10.7b_3.0.Q2_K.gguf	Q2_K	3.73GB
Myrrh_solar_10.7b_3.0.IQ3_XS.gguf	IQ3_XS	4.14GB
Myrrh_solar_10.7b_3.0.IQ3_S.gguf	IQ3_S	4.37GB
Myrrh_solar_10.7b_3.0.Q3_K_S.gguf	Q3_K_S	4.34GB
Myrrh_solar_10.7b_3.0.IQ3_M.gguf	IQ3_M	4.51GB
Myrrh_solar_10.7b_3.0.Q3_K.gguf	Q3_K	4.84GB
Myrrh_solar_10.7b_3.0.Q3_K_M.gguf	Q3_K_M	4.84GB
Myrrh_solar_10.7b_3.0.Q3_K_L.gguf	Q3_K_L	5.26GB
Myrrh_solar_10.7b_3.0.IQ4_XS.gguf	IQ4_XS	5.43GB
Myrrh_solar_10.7b_3.0.Q4_0.gguf	Q4_0	5.66GB
Myrrh_solar_10.7b_3.0.IQ4_NL.gguf	IQ4_NL	5.72GB
Myrrh_solar_10.7b_3.0.Q4_K_S.gguf	Q4_K_S	5.7GB
Myrrh_solar_10.7b_3.0.Q4_K.gguf	Q4_K	6.02GB
Myrrh_solar_10.7b_3.0.Q4_K_M.gguf	Q4_K_M	6.02GB
Myrrh_solar_10.7b_3.0.Q4_1.gguf	Q4_1	6.27GB
Myrrh_solar_10.7b_3.0.Q5_0.gguf	Q5_0	6.89GB
Myrrh_solar_10.7b_3.0.Q5_K_S.gguf	Q5_K_S	6.89GB
Myrrh_solar_10.7b_3.0.Q5_K.gguf	Q5_K	7.08GB
Myrrh_solar_10.7b_3.0.Q5_K_M.gguf	Q5_K_M	7.08GB
Myrrh_solar_10.7b_3.0.Q5_1.gguf	Q5_1	7.51GB
Myrrh_solar_10.7b_3.0.Q6_K.gguf	Q6_K	8.2GB
Myrrh_solar_10.7b_3.0.Q8_0.gguf	Q8_0	10.62GB

Original model description:

license: apache-2.0 language: - ko

Model Details

Model Developers : Taeeon Park, Gihong Lee

dataset : dpo medical dataset (AI-hub dataset 활용 자체 제작)

Training Method Method : DPO.

Company : MoAData

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
repo = "MoaData/Myrrh_solar_10.7b_3.0"
model = AutoModelForCausalLM.from_pretrained(
        repo,
        return_dict=True,
        torch_dtype=torch.float16,
        device_map='auto'
)
tokenizer = AutoTokenizer.from_pretrained(repo)