CoolFace
Modelpublic

RichardErkhov/dddsaty_-_FusionNet_7Bx2_MoE_Ko_DPO_Adapter_Attach-gguf

sourceHugging Faceupdated 2y agoView on Hugging Face
0likes427downloads
Model Card

Quantization made by Richard Erkhov.

Github

Discord

Request more models

FusionNet7Bx2MoEKoDPOAdapterAttach - GGUF

  • —Model creator: https://huggingface.co/dddsaty/
  • —Original model: https://huggingface.co/dddsaty/FusionNet7Bx2MoEKoDPOAdapterAttach/
NameQuant methodSize
FusionNet_7Bx2_MoE_Ko_DPO_Adapter_Attach.Q2_K.ggufQ2_K4.43GB
FusionNet_7Bx2_MoE_Ko_DPO_Adapter_Attach.IQ3_XS.ggufIQ3_XS4.95GB
FusionNet_7Bx2_MoE_Ko_DPO_Adapter_Attach.IQ3_S.ggufIQ3_S5.22GB
FusionNet_7Bx2_MoE_Ko_DPO_Adapter_Attach.Q3_K_S.ggufQ3KS5.2GB
FusionNet_7Bx2_MoE_Ko_DPO_Adapter_Attach.IQ3_M.ggufIQ3_M5.35GB
FusionNet_7Bx2_MoE_Ko_DPO_Adapter_Attach.Q3_K.ggufQ3_K5.78GB
FusionNet_7Bx2_MoE_Ko_DPO_Adapter_Attach.Q3_K_M.ggufQ3KM5.78GB
FusionNet_7Bx2_MoE_Ko_DPO_Adapter_Attach.Q3_K_L.ggufQ3KL6.27GB
FusionNet_7Bx2_MoE_Ko_DPO_Adapter_Attach.IQ4_XS.ggufIQ4_XS6.5GB
FusionNet_7Bx2_MoE_Ko_DPO_Adapter_Attach.Q4_0.ggufQ4_06.78GB
FusionNet_7Bx2_MoE_Ko_DPO_Adapter_Attach.IQ4_NL.ggufIQ4_NL6.85GB
FusionNet_7Bx2_MoE_Ko_DPO_Adapter_Attach.Q4_K_S.ggufQ4KS6.84GB
FusionNet_7Bx2_MoE_Ko_DPO_Adapter_Attach.Q4_K.ggufQ4_K7.25GB
FusionNet_7Bx2_MoE_Ko_DPO_Adapter_Attach.Q4_K_M.ggufQ4KM7.25GB
FusionNet_7Bx2_MoE_Ko_DPO_Adapter_Attach.Q4_1.ggufQ4_17.52GB
FusionNet_7Bx2_MoE_Ko_DPO_Adapter_Attach.Q5_0.ggufQ5_08.26GB
FusionNet_7Bx2_MoE_Ko_DPO_Adapter_Attach.Q5_K_S.ggufQ5KS8.26GB
FusionNet_7Bx2_MoE_Ko_DPO_Adapter_Attach.Q5_K.ggufQ5_K8.51GB
FusionNet_7Bx2_MoE_Ko_DPO_Adapter_Attach.Q5_K_M.ggufQ5KM8.51GB
FusionNet_7Bx2_MoE_Ko_DPO_Adapter_Attach.Q5_1.ggufQ5_19.01GB
FusionNet_7Bx2_MoE_Ko_DPO_Adapter_Attach.Q6_K.ggufQ6_K9.84GB
FusionNet_7Bx2_MoE_Ko_DPO_Adapter_Attach.Q8_0.ggufQ8_012.75GB

Original model description: --- license: mit datasets: We-Want-GPU/Yi-Ko-DPO-Orca-DPO-Pairs language:

  • —en
  • —ko pipeline_tag: text-generation ---

Explanation

  • —With the base model, attached the DPO applied Adapter

Base Model

Adapter Base Model

Adapter Corpus

Score |Average|ARC|HellaSwag|MMLU|TruthfulQA|Winogrande|GSM8K| |:---:|:---:|:---:|:---:|:---:|:---:|:---:| |76.09|73.89|88.94|65.03|71.24|87.61|69.83|

Log

  • —2024.02.13: Initial version Upload

LICENSE

  • —MIT