metadata
base_model:
- unsloth/DeepSeek-R1-Distill-Qwen-1.5B-bnb-4bit
tags:
- text-generation-inference
- transformers
- unsloth
- qwen2
- trl
license: apache-2.0
language:
- en
- zh
datasets:
- FradSer/DeepSeek-R1-Distilled-Translate-en-zh_CN-39k
- wmt/wmt19
pipeline_tag: text-generation
DeepTranslate-R1-1.5B 概述
DeepTranslate-R1-1.5B 是一个从 DeepSeek-R1-Distilled-Qwen-1.5B 微调而来的语言模型，专门用于英文和中文之间的高质量翻译。我们的模型使用监督式微调(SFT)技术，在仅有1.5B参数的计算效率下实现高质量翻译。
输出模板
TEMPLATE """{{- if .System }}{{ .System }}{{ end }}
{{- range $i, $_ := .Messages }}
{{- $last := eq (len (slice $.Messages $i)) 1}}
{{- if eq .Role "user" }}<｜User｜>{{ .Content }}
{{- else if eq .Role "assistant" }}<｜Assistant｜>{{ .Content }}{{- if not $last }}<｜end▁of▁sentence｜>{{- end }}
{{- end }}
{{- if and $last (ne .Role "assistant") }}<｜Assistant｜>{{- end }}
{{- end }}"""
作者
This qwen2 model was trained 2x faster with Unsloth and Huggingface's TRL library.
