Modelpublic

RichardErkhov/hivaze_-_ParaLex-Llama-3-8B-SFT-gguf

sourceHugging Faceupdated 2y agoView on Hugging Face

0likes349downloads

Model Card

Quantization made by Richard Erkhov.

Github

Discord

Request more models

ParaLex-Llama-3-8B-SFT - GGUF

—Model creator: https://huggingface.co/hivaze/
—Original model: https://huggingface.co/hivaze/ParaLex-Llama-3-8B-SFT/

Name	Quant method	Size
ParaLex-Llama-3-8B-SFT.Q2_K.gguf	Q2_K	2.96GB
ParaLex-Llama-3-8B-SFT.IQ3_XS.gguf	IQ3_XS	3.28GB
ParaLex-Llama-3-8B-SFT.IQ3_S.gguf	IQ3_S	3.43GB
ParaLex-Llama-3-8B-SFT.Q3_K_S.gguf	Q3KS	3.41GB
ParaLex-Llama-3-8B-SFT.IQ3_M.gguf	IQ3_M	3.52GB
ParaLex-Llama-3-8B-SFT.Q3_K.gguf	Q3_K	3.74GB
ParaLex-Llama-3-8B-SFT.Q3_K_M.gguf	Q3KM	3.74GB
ParaLex-Llama-3-8B-SFT.Q3_K_L.gguf	Q3KL	4.03GB
ParaLex-Llama-3-8B-SFT.IQ4_XS.gguf	IQ4_XS	4.18GB
ParaLex-Llama-3-8B-SFT.Q4_0.gguf	Q4_0	4.34GB
ParaLex-Llama-3-8B-SFT.IQ4_NL.gguf	IQ4_NL	4.38GB
ParaLex-Llama-3-8B-SFT.Q4_K_S.gguf	Q4KS	4.37GB
ParaLex-Llama-3-8B-SFT.Q4_K.gguf	Q4_K	4.58GB
ParaLex-Llama-3-8B-SFT.Q4_K_M.gguf	Q4KM	4.58GB
ParaLex-Llama-3-8B-SFT.Q4_1.gguf	Q4_1	4.78GB
ParaLex-Llama-3-8B-SFT.Q5_0.gguf	Q5_0	5.21GB
ParaLex-Llama-3-8B-SFT.Q5_K_S.gguf	Q5KS	5.21GB
ParaLex-Llama-3-8B-SFT.Q5_K.gguf	Q5_K	5.34GB
ParaLex-Llama-3-8B-SFT.Q5_K_M.gguf	Q5KM	5.34GB
ParaLex-Llama-3-8B-SFT.Q5_1.gguf	Q5_1	5.65GB
ParaLex-Llama-3-8B-SFT.Q6_K.gguf	Q6_K	6.14GB
ParaLex-Llama-3-8B-SFT.Q8_0.gguf	Q8_0	7.95GB

Original model description: --- language:

—en
—ru ---

Russified version of Llama-3-8b-Instruct

The model was obtained as a result of translation of the English answers of the original model in Russian and additional training in English and Russian.

No responses from third-party models (GPT-4, etc.) were used, only from the original one, so the model must maintain alignment.