llama3

Meta Llama 3：迄今为止最强大的公开可用 LLM

8b 70b

6.6M 拉取更新 5 个月前

68 个标签

更新于 5 个月前

5 个月前

4c1a815173f7 · 31GB

{"num_keep":24,"stop":["\u003c|start_header_id|\u003e","\u003c|end_header_id|\u003e","\u003c|eot_id|

110B

模板

{{ if .System }}<|start_header_id|>system<|end_header_id|> {{ .System }}<|eot_id|>{{ end }}{{ if .P

254B

许可证

META LLAMA 3 COMMUNITY LICENSE AGREEMENT Meta Llama 3 Version Release Date: April 18, 2024 “Agree

12kB

自述文件

Llama 3

迄今为止最强大的公开可用 LLM。

Meta Llama 3 是 Meta Inc. 开发的一系列模型，它们是新的最先进模型，有 **8B** 和 **70B** 两种参数大小（预训练或指令微调）。

Llama 3 指令微调模型经过微调和优化，适用于对话/聊天用例，在常见基准测试中优于许多可用的开源聊天模型。

CLI

打开终端并运行 ollama run llama3

API

使用 curl 的示例

curl -X POST https://127.0.0.1:11434/api/generate -d '{
  "model": "llama3",
  "prompt":"Why is the sky blue?"
 }'

API 文档

模型变体

Instruct 针对聊天/对话用例进行了微调。

示例： ollama run llama3 ollama run llama3:70b

预训练 是基础模型。

示例： ollama run llama3:text ollama run llama3:70b-text

参考

Introducing Meta Llama 3: The most capable openly available LLM to date

# Llama 3

The most capable openly available LLM to date.

Meta Llama 3, a family of models developed by Meta Inc. are new state-of-the-art , available in both **8B** and **70B** parameter sizes (pre-trained or instruction-tuned).

Llama 3 instruction-tuned models are fine-tuned and optimized for dialogue/chat use cases and outperform many of the available open-source chat models on common benchmarks.

### CLI

Open the terminal and run `ollama run llama3`

### API

Example using curl:

```bash
curl -X POST https://127.0.0.1:11434/api/generate -d '{
  "model": "llama3",
  "prompt":"Why is the sky blue?"
 }'
```

[API documentation](https://github.com/ollama/ollama/blob/main/docs/api.md)

## Model variants

**Instruct** is fine-tuned for chat/dialogue use cases.

*Example:*
`ollama run llama3`
`ollama run llama3:70b`

**Pre-trained** is the base model.

*Example:*
`ollama run llama3:text`
`ollama run llama3:70b-text`

## References
[Introducing Meta Llama 3: The most capable openly available LLM to date](https://ai.meta.com/blog/meta-llama-3/)

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