falcon:40b-instruct

falcon

Falcon 是由技术创新研究所 (TII) 构建的大型语言模型，用于摘要、文本生成和聊天机器人。

7b 40b 180b

67.2K 下载量更新于 15 个月前

38 个标签

更新于 15 个月前

15 个月前

bc9368437a24 · 24GB

{ "stop": [ "User:", "Assistant:" ] }

31B

模板

{{ .System }} User: {{ .Prompt }} Assistant:

45B

自述文件

Falcon 是由技术创新研究所 (TII) 构建的一系列高性能大型语言模型，TII 是阿布扎比政府高级技术研究委员会下属的研究中心，负责监管技术研究。

命令行界面

ollama run falcon "Why is the sky blue?"

应用程序接口

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

参数计数

参数计数	推荐内存
70 亿	8GB	查看	`ollama run falcon:7b`
400 亿	32GB	查看	`ollama run falcon:40b`
1800 亿	192GB	查看	`ollama run falcon:180b`

变体


`聊天`	聊天模型在聊天和指令数据集上进行了微调，其中混合了多个大型对话数据集。
`指令`	指令模型遵循指令，并在 baize 指令数据集上进行了微调。
`文本`	文本模型是没有任何对话微调的基础模型，最适合用于简单的文本补全。

Falcon 180B

截至 2023 年 9 月，拥有 1800 亿参数的模型 Falcon 180B 是性能最佳的公开 LLM。它的性能介于 OpenAI 的 GPT 3.5 和 GPT 4 之间。为了运行 Falcon 180B，建议使用至少 192GB 总内存的强大系统。

注意：Falcon 180B 的发布许可与其较小的同级产品不同，该许可限制在某些条件下的商业用途。有关更多信息，请参阅模型详情和许可。

更多信息

Falcon is a family of high-performing large language models model built by the Technology Innovation Institute (TII), a research center part of Abu Dhabi government’s advanced technology research council overseeing technology research.

### CLI

```
ollama run falcon "Why is the sky blue?"
```

### API

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

## Parameter counts
| Parameter Count | Recommended memory |                              |                          |
| --------------- | ------------------ | ---------------------------- | ------------------------ |
| 7 billion       | 8GB               | [View](/library/falcon:7b)   | `ollama run falcon:7b`   |
| 40 billion      | 32GB               | [View](/library/falcon:40b)  | `ollama run falcon:40b`  |
| 180 billion     | 192GB              | [View](/library/falcon:180b) | `ollama run falcon:180b` |

## Variations

|            |                                                                                                                                    |
| ---------- | ---------------------------------------------------------------------------------------------------------------------------------- |
| `chat`     | Chat models are fine-tuned on chat and instruction datasets with a mix of several large-scale conversational datasets.             |
| `instruct` | Instruct models follow instructions and are fine-tuned on the [baize](https://www.google.com/search?q=baize+dataset&oq=baize+data&aqs=chrome.0.0i512j69i57j0i10i15i22i30i625j0i390i650.1387j0j7&sourceid=chrome&ie=UTF-8) instructional dataset.                                                     |
| `text`     | Text models are the base foundation model without any fine-tuning for conversations, and are best used for simple text completion. |

## Falcon 180B

As of September 2023, the 180 billion parameter model, Falcon 180B, is the best-performing openly released LLM. It sits somewhere in between OpenAI's GPT 3.5 and GPT 4. For running Falcon 180B, a powerful system is recommended with at least 192GB of total memory.

> Note: Falcon 180B is released under a different license than its smaller siblings that restricts commercial use under certain conditions. See the [model details](/library/falcon:180b) and license for more information.

## More information

* [TII's website](https://www.tii.ae/)
* [Falcon 180B announcement](https://falconllm.tii.ae)
* [TII on HuggingFace](http://hugging-face.cn/tiiuae)

粘贴、拖放或点击以上传图片 (.png, .jpeg, .jpg, .svg, .gif)