> 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-de/training/training.md).

# Überblick

KI-Modelle auf CLORE.AI-GPUs trainieren und feinabstimmen.

## Verfügbare Leitfäden

| Anleitung                                                                   | Anwendungsfall                          | Schwierigkeitsgrad |
| --------------------------------------------------------------------------- | --------------------------------------- | ------------------ |
| [Jupyter ML Training](/guides/guides_v2-de/training/jupyter-ml-training.md) | Interaktives Training                   | Einfach            |
| [DreamBooth](/guides/guides_v2-de/training/dreambooth.md)                   | Benutzerdefinierte SD-Subjekte          | Mittel             |
| [Kohya Training](/guides/guides_v2-de/training/kohya-training.md)           | LoRA/LyCORIS-Training                   | Mittel             |
| [LLM feinabstimmen](/guides/guides_v2-de/training/finetune-llm.md)          | LLM-Feinabstimmung                      | Fortgeschritten    |
| [DeepSpeed](/guides/guides_v2-de/training/deepspeed-training.md)            | Verteiltes Training                     | Fortgeschritten    |
| [HuggingFace](/guides/guides_v2-de/training/huggingface-transformers.md)    | Transformers-Training                   | Mittel             |
| [Unsloth](/guides/guides_v2-de/training/unsloth-finetune.md)                | Schnelle LLM-Feinabstimmung             | Mittel             |
| [Axolotl](/guides/guides_v2-de/training/axolotl-training.md)                | YAML-first Feinabstimmung               | Mittel             |
| [LLaMA-Factory](/guides/guides_v2-de/training/llama-factory.md)             | Einfaches LLM-Training UI               | Einfach            |
| [TRL](/guides/guides_v2-de/training/trl.md)                                 | Transformer-Verstärkungslernen          | Fortgeschritten    |
| [LitGPT](/guides/guides_v2-de/training/litgpt.md)                           | Lightning-basiertes Training            | Mittel             |
| [Mergekit](/guides/guides_v2-de/training/mergekit.md)                       | Toolkit zum Zusammenführen von Modellen | Einfach            |

## GPU-Empfehlungen

| Aufgabe                  | Minimum  | Empfohlen |
| ------------------------ | -------- | --------- |
| LoRA (SD)                | RTX 3060 | RTX 3090  |
| DreamBooth               | RTX 3090 | RTX 4090  |
| LLM-Feinabstimmung (7B)  | RTX 3090 | A100 40GB |
| LLM-Feinabstimmung (70B) | 4x A100  | 8x A100   |

## Trainingstypen

### Bildmodelle

* **LoRA** - Leichter Adapter, schnelles Training
* **DreamBooth** - Vollständige Feinabstimmung für Konzepte
* **Textual Inversion** - Neue Tokens lernen

### Sprachmodelle

* **LoRA/QLoRA** - Speicherplatzsparende Feinabstimmung
* **Volle Feinabstimmung** - Beste Qualität, benötigt mehr VRAM

## Tipps

* Verwendung **Spot** Bestellungen für lange Trainingsläufe
* Gradient Checkpointing aktivieren, um VRAM zu sparen
* Training mit TensorBoard überwachen
* Checkpoints häufig speichern

## Verwandte Anleitungen

* [Sprachmodelle](/guides/guides_v2-de/sprachmodelle/language-models.md)
* [Bildgenerierung](/guides/guides_v2-de/bildgenerierung/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:

```
GET https://docs.clore.ai/guides/guides_v2-de/training/training.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
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.