orca-mini

Orca Mini 是一个在 Orca 风格的数据集上训练的 Llama 和 Llama 2 模型，这些数据集使用论文中定义的方法创建，论文名称为 Orca: Progressive Learning from Complex Explanation Traces of GPT-4。有两种变体可用。基于 Llama 的原始 Orca Mini，参数量为 30 亿、70 亿和 130 亿，以及基于 Llama 2 的 v3 版本，参数量为 70 亿、130 亿和 700 亿。

用法

CLI

打开终端并运行 ollama run orca-mini

API

示例

  curl -X POST https://127.0.0.1:11434/api/generate -d '{
    "model": "orca-mini",
    "prompt":"Why is the sky blue?"
   }'

内存需求

7b 模型通常需要至少 8GB 的内存
13b 模型通常需要至少 16GB 的内存
70b 模型通常需要至少 64GB 的内存

参考

3b 参数原始来源：Pankaj Mathur

7b 参数原始来源：Pankaj Mathur

13b 参数原始来源：Pankaj Mathur

Ollama 上的 Orca Mini v3 源代码

13b 参数原始来源：Pankaj Mathur

70b 参数来源：Pankaj Mathur

Orca: Progressive Learning from Complex Explanation Traces of GPT-4

Orca Mini is a Llama and Llama 2 model trained on Orca Style datasets created using the approaches defined in the paper, Orca: Progressive Learning from Complex Explanation Traces of GPT-4. There are two variations available. The original Orca Mini based on Llama in 3, 7, and 13 billion parameter sizes, and v3 based on Llama 2 in 7, 13, and 70 billion parameter sizes.

## Usage

### CLI

Open the terminal and run `ollama run orca-mini`

### API

Example:

```bash
  curl -X POST https://127.0.0.1:11434/api/generate -d '{
    "model": "orca-mini",
    "prompt":"Why is the sky blue?"
   }'
  ```

## Memory requirements

- 7b models generally require at least 8GB of RAM
- 13b models generally require at least 16GB of RAM
- 70b models generally require at least 64GB of RAM

## Reference

3b parameters original source:
 [Pankaj Mathur](https://hugging-face.cn/psmathur/orca_mini_3b)

7b parameters original source:
 [Pankaj Mathur](https://hugging-face.cn/psmathur/orca_mini_7b)

13b parameters original source:
 [Pankaj Mathur](https://hugging-face.cn/psmathur/orca_mini_13b)

**Orca Mini v3 source on Ollama**

13b parameters original source:
 [Pankaj Mathur](https://hugging-face.cn/psmathur/orca_mini_v3_13b)

70b parameters source:
 [Pankaj Mathur](https://hugging-face.cn/psmathur/orca_mini_v3_70b)

[Orca: Progressive Learning from Complex Explanation Traces of GPT-4](https://arxiv.org/abs/2306.02707 )

粘贴、拖放或点击上传图像（.png、.jpeg、.jpg、.svg、.gif）