Modelpublic

RichardErkhov/TriadParty_-_deepsex-6b-base-gguf

sourceHugging Faceupdated 2y agoView on Hugging Face

0likes104downloads

Model Card

Quantization made by Richard Erkhov.

Request more models

deepsex-6b-base - GGUF

—Model creator: https://huggingface.co/TriadParty/
—Original model: https://huggingface.co/TriadParty/deepsex-6b-base/

Name	Quant method	Size
deepsex-6b-base.Q2_K.gguf	Q2_K	2.18GB
deepsex-6b-base.IQ3_XS.gguf	IQ3_XS	2.41GB
deepsex-6b-base.IQ3_S.gguf	IQ3_S	2.53GB
deepsex-6b-base.Q3_K_S.gguf	Q3KS	2.52GB
deepsex-6b-base.IQ3_M.gguf	IQ3_M	2.62GB
deepsex-6b-base.Q3_K.gguf	Q3_K	2.79GB
deepsex-6b-base.Q3_K_M.gguf	Q3KM	2.79GB
deepsex-6b-base.Q3_K_L.gguf	Q3KL	3.01GB
deepsex-6b-base.IQ4_XS.gguf	IQ4_XS	3.11GB
deepsex-6b-base.Q4_0.gguf	Q4_0	3.24GB
deepsex-6b-base.IQ4_NL.gguf	IQ4_NL	3.27GB
deepsex-6b-base.Q4_K_S.gguf	Q4KS	3.26GB
deepsex-6b-base.Q4_K.gguf	Q4_K	3.42GB
deepsex-6b-base.Q4_K_M.gguf	Q4KM	3.42GB
deepsex-6b-base.Q4_1.gguf	Q4_1	3.58GB
deepsex-6b-base.Q5_0.gguf	Q5_0	3.92GB
deepsex-6b-base.Q5_K_S.gguf	Q5KS	3.92GB
deepsex-6b-base.Q5_K.gguf	Q5_K	4.01GB
deepsex-6b-base.Q5_K_M.gguf	Q5KM	4.01GB
deepsex-6b-base.Q5_1.gguf	Q5_1	4.25GB
deepsex-6b-base.Q6_K.gguf	Q6_K	4.63GB
deepsex-6b-base.Q8_0.gguf	Q8_0	6.0GB

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

—not-for-all-audiences ---

LUST series

For the 34b version, see this

Compared with 34b, the main changes are:

1.The base model uses the YI-6B version modified to the llama layer structure, which should have better adaptability to many applications.
2.Use a larger data set
3.This is a base model. You can use sft+dpo to train a model that is more suitable for you.
4.part of dataset is available in deepsex-RP