RichardErkhov/Jaehyeon222_-_ME-MOE-7Bx2_test-gguf

Quantization made by Richard Erkhov.

Github

Discord

Request more models

ME-MOE-7Bx2_test - GGUF

• Model creator: https://huggingface.co/Jaehyeon222/
• Original model: https://huggingface.co/Jaehyeon222/ME-MOE-7Bx2_test/

Name	Quant method	Size
ME-MOE-7Bx2_test.Q2_K.gguf	Q2_K	4.43GB
ME-MOE-7Bx2_test.IQ3_XS.gguf	IQ3_XS	4.95GB
ME-MOE-7Bx2_test.IQ3_S.gguf	IQ3_S	5.22GB
ME-MOE-7Bx2_test.Q3_K_S.gguf	Q3_K_S	5.2GB
ME-MOE-7Bx2_test.IQ3_M.gguf	IQ3_M	5.35GB
ME-MOE-7Bx2_test.Q3_K.gguf	Q3_K	5.78GB
ME-MOE-7Bx2_test.Q3_K_M.gguf	Q3_K_M	5.78GB
ME-MOE-7Bx2_test.Q3_K_L.gguf	Q3_K_L	6.27GB
ME-MOE-7Bx2_test.IQ4_XS.gguf	IQ4_XS	6.5GB
ME-MOE-7Bx2_test.Q4_0.gguf	Q4_0	6.78GB
ME-MOE-7Bx2_test.IQ4_NL.gguf	IQ4_NL	6.85GB
ME-MOE-7Bx2_test.Q4_K_S.gguf	Q4_K_S	6.84GB
ME-MOE-7Bx2_test.Q4_K.gguf	Q4_K	7.25GB
ME-MOE-7Bx2_test.Q4_K_M.gguf	Q4_K_M	7.25GB
ME-MOE-7Bx2_test.Q4_1.gguf	Q4_1	7.52GB
ME-MOE-7Bx2_test.Q5_0.gguf	Q5_0	8.26GB
ME-MOE-7Bx2_test.Q5_K_S.gguf	Q5_K_S	8.26GB
ME-MOE-7Bx2_test.Q5_K.gguf	Q5_K	8.51GB
ME-MOE-7Bx2_test.Q5_K_M.gguf	Q5_K_M	8.51GB
ME-MOE-7Bx2_test.Q5_1.gguf	Q5_1	9.01GB
ME-MOE-7Bx2_test.Q6_K.gguf	Q6_K	9.84GB
ME-MOE-7Bx2_test.Q8_0.gguf	Q8_0	12.75GB

Original model description:

license: cc-by-nc-nd-4.0 language: - ko pipeline_tag: text-generation tags: - Mixture of experts

Model Card for ME-MOE-7Bx2_test

Developed by : 메가스터디교육, 프리딕션, 마이스

Base Model : megastudyedu/ME-dpo-7B-v1.0

Expert Models : megastudyedu/ME-dpo-7B-v1.1, macadeliccc/WestLake-7B-v2-laser-truthy-dpo

Method : merge-kit을 활용하여 MOE를 구현했습니다.

---

ME-7B-v1.0 -> ME-dpo-7B-v1.0, ME-7B-v1.1 -> ME-dpo-7B-v1.1 로 수정했습니다.