eramax/nxcode-cq-7b-orpo:q6

eramax/

nxcode-cq-7b-orpo:q6

1,875 Downloads Updated 1 year ago

https://huggingface.co/NTQAI/Nxcode-CQ-7B-orpo

Updated 1 year ago

1 year ago

2784da3b3724 · 6.4GB ·

archqwen2

parameters7.25B

quantizationQ6_K

6.4GB

{ "stop": [ "<|im_start|>", "<|im_end|>", "</s>" ] }

76B

{{ if .System }}<|im_start|>system {{ .System }}<|im_end|> {{ end }}{{ if .Prompt }}<|im_start|>user

186B

Nxcode-CQ-7B-orpo is an Monolithic Preference Optimization without Reference Model fine-tune of Qwen/CodeQwen1.5-7B on 100k samples of high-quality ranking data.