Modelpublic

RichardErkhov/etri-xainlp_-_llama2-13b-sft-dpo-gguf

sourceHugging Faceupdated 2y agoView on Hugging Face

0likes68downloads

Model Card

Quantization made by Richard Erkhov.

Request more models

llama2-13b-sft-dpo - GGUF

—Model creator: https://huggingface.co/etri-xainlp/
—Original model: https://huggingface.co/etri-xainlp/llama2-13b-sft-dpo/

Name	Quant method	Size
llama2-13b-sft-dpo.Q2_K.gguf	Q2_K	4.52GB
llama2-13b-sft-dpo.IQ3_XS.gguf	IQ3_XS	4.99GB
llama2-13b-sft-dpo.IQ3_S.gguf	IQ3_S	5.27GB
llama2-13b-sft-dpo.Q3_K_S.gguf	Q3KS	5.27GB
llama2-13b-sft-dpo.IQ3_M.gguf	IQ3_M	5.57GB
llama2-13b-sft-dpo.Q3_K.gguf	Q3_K	5.9GB
llama2-13b-sft-dpo.Q3_K_M.gguf	Q3KM	5.9GB
llama2-13b-sft-dpo.Q3_K_L.gguf	Q3KL	6.45GB
llama2-13b-sft-dpo.IQ4_XS.gguf	IQ4_XS	6.54GB
llama2-13b-sft-dpo.Q4_0.gguf	Q4_0	6.86GB
llama2-13b-sft-dpo.IQ4_NL.gguf	IQ4_NL	6.9GB
llama2-13b-sft-dpo.Q4_K_S.gguf	Q4KS	6.91GB
llama2-13b-sft-dpo.Q4_K.gguf	Q4_K	7.33GB
llama2-13b-sft-dpo.Q4_K_M.gguf	Q4KM	7.33GB
llama2-13b-sft-dpo.Q4_1.gguf	Q4_1	7.61GB
llama2-13b-sft-dpo.Q5_0.gguf	Q5_0	8.36GB
llama2-13b-sft-dpo.Q5_K_S.gguf	Q5KS	8.36GB
llama2-13b-sft-dpo.Q5_K.gguf	Q5_K	8.6GB
llama2-13b-sft-dpo.Q5_K_M.gguf	Q5KM	8.6GB
llama2-13b-sft-dpo.Q5_1.gguf	Q5_1	9.1GB
llama2-13b-sft-dpo.Q6_K.gguf	Q6_K	9.95GB
llama2-13b-sft-dpo.Q8_0.gguf	Q8_0	12.88GB

Original model description: --- license: apache-2.0 ---

etri-xainlp/llama2-13b-sft-dpo

Model Details

Model Developers ETRI xainlp team

Input text only.

Output text only.

Model Architecture

Base Model meta-llama/Llama-13b-hf

Training Dataset

—fully sft: 650k instruction-following set

—dpo+lora: 90k user preference set

—We use A100 GPU 80GB * 8, when training.