chat application optimization, fp16 architecture, zephyr model adaptation

Updated 2 years ago

2 years ago

d07fbe156d74 · 14GB ·

archllama

parameters7.24B

quantizationF16

14GB

{{- if .First }} ### System: {{ .System }} {{- end }} ### User: {{ .Prompt }} ### Response:

95B

{ "stop": [ "<|system|>", "<|user|>", "<|assistant|>", "</s>"

97B

Readme

This is an adapted version of the original Zephyr model, based on the FP16 architecture and specifically tailored for chat applications.