ima/
deepseek-math:latest

262 1 year ago

1 year ago

088043480091 · 4.0GB

llama
·
6.91B
·
Q4_0
{{ if .System }}{{ .System }} {{ end }}{{ if .Prompt }}User: {{ .Prompt }} {{ end }}Assistant: {{ .R

Readme

Unofficial upload of the deepseek-math model to ollama. Source: https://huggingface.co/deepseek-ai/deepseek-math-7b-rl

It has been quantitized to q4_0.

Tips:

  • As the authors have stated, add the following line of text after your question for better responses: “Please reason step by step, and put your final answer within \boxed{}.”
  • This model is suited only for answering math questions. It will not converse or discuss anything else.
  • Don’t ask multiple questions in the same conversation thread. Create a new thread for each question.
  • The model does not have vision capabilities
  • You can input formulas in large diversity of ways, it will understand almost any format