> For the complete documentation index, see [llms.txt](https://docs.clore.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.clore.ai/guides/guides_v2-zh/xun-lian/training.md).

# 概览

在 CLORE.AI GPU 上训练和微调 AI 模型。

## 可用指南

| 指南                                                                       | 使用场景             | 难度 |
| ------------------------------------------------------------------------ | ---------------- | -- |
| [Jupyter 机器学习训练](/guides/guides_v2-zh/xun-lian/jupyter-ml-training.md)   | 交互式训练            | 简单 |
| [DreamBooth](/guides/guides_v2-zh/xun-lian/dreambooth.md)                | 自定义 SD 主题        | 中等 |
| [Kohya 训练](/guides/guides_v2-zh/xun-lian/kohya-training.md)              | LoRA/LyCORIS 训练  | 中等 |
| [微调 LLM](/guides/guides_v2-zh/xun-lian/finetune-llm.md)                  | LLM 微调           | 高级 |
| [DeepSpeed](/guides/guides_v2-zh/xun-lian/deepspeed-training.md)         | 分布式训练            | 高级 |
| [HuggingFace](/guides/guides_v2-zh/xun-lian/huggingface-transformers.md) | Transformers 训练  | 中等 |
| [Unsloth](/guides/guides_v2-zh/xun-lian/unsloth-finetune.md)             | 快速 LLM 微调        | 中等 |
| [Axolotl](/guides/guides_v2-zh/xun-lian/axolotl-training.md)             | 以 YAML 为先的微调     | 中等 |
| [LLaMA-Factory](/guides/guides_v2-zh/xun-lian/llama-factory.md)          | 简易 LLM 训练界面      | 简单 |
| [TRL](/guides/guides_v2-zh/xun-lian/trl.md)                              | Transformer 强化学习 | 高级 |
| [LitGPT](/guides/guides_v2-zh/xun-lian/litgpt.md)                        | 基于 Lightning 的训练 | 中等 |
| [Mergekit](/guides/guides_v2-zh/xun-lian/mergekit.md)                    | 模型合并工具包          | 简单 |

## GPU 推荐

{% hint style="warning" %}
**Clore.ai 市场上未列出多 GPU 的 80GB 级机型。** 目前列出的最大配置是 4× RTX PRO 6000 Blackwell（每张 96GB，共 380GB）以及 8–11× RTX 5090（每张 32GB）。A100 / H200 / B200 容量可按 [裸机](https://clore.ai/bare-metal) 需求提供。部署前请查看 [GPU 价格与可用性](/guides/guides_v2-zh/ru-men-zhi-nan/pricing.md) 。
{% endhint %}

| 任务          | 最低       | 推荐        |
| ----------- | -------- | --------- |
| LoRA（SD）    | RTX 3060 | RTX 3090  |
| DreamBooth  | RTX 3090 | RTX 4090  |
| LLM 微调（7B）  | RTX 3090 | A100 40GB |
| LLM 微调（70B） | 4 张 A100 | 8 张 A100  |

## 训练类型

### 图像模型

* **LoRA** - 轻量级适配器，训练快速
* **DreamBooth** - 用于概念的全量微调
* **文本反演** - 学习新 token

### 语言模型

* **LoRA/QLoRA** - 内存高效微调
* **全量微调** - 质量最佳，但需要更多显存

## 提示

* 使用 **竞价** 长时间训练运行的订单
* 启用梯度检查点以节省显存
* 使用 TensorBoard 监控训练
* 经常保存检查点

## 相关指南

* [语言模型](/guides/guides_v2-zh/yu-yan-mo-xing/language-models.md)
* [图像生成](/guides/guides_v2-zh/tu-xiang-sheng-cheng/image-generation.md)


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.clore.ai/guides/guides_v2-zh/xun-lian/training.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.