专注于数学和逻辑问题的模型

7b 13b 70b

101.1K 14 个月前

自述文件

现在已更新至 WizardMath 7B v1.1: ollama pull wizard-math。这个新版本使用 Mistral-7B 训练,并且比以前的版本获得了更高的基准分数。

WizardMath 由 WizardLM 发布。它在 GSM8k 数据集上训练,并针对数学问题。它有 7B、13B 和 70B 参数大小可供选择。

示例提示

How many 4-digit numbers have the last digit equal to the sum of the first two digits?

参考

HuggingFace