Instructions to use Keyven/german-ocr-2b-gguf with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Keyven/german-ocr-2b-gguf with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="Keyven/german-ocr-2b-gguf",
	filename="German-OCR-Engine.2B.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": [
				{
					"type": "text",
					"text": "Describe this image in one sentence."
				},
				{
					"type": "image_url",
					"image_url": {
						"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
					}
				}
			]
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use Keyven/german-ocr-2b-gguf with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf Keyven/german-ocr-2b-gguf
# Run inference directly in the terminal:
llama-cli -hf Keyven/german-ocr-2b-gguf

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf Keyven/german-ocr-2b-gguf
# Run inference directly in the terminal:
llama-cli -hf Keyven/german-ocr-2b-gguf

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf Keyven/german-ocr-2b-gguf
# Run inference directly in the terminal:
./llama-cli -hf Keyven/german-ocr-2b-gguf

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf Keyven/german-ocr-2b-gguf
# Run inference directly in the terminal:
./build/bin/llama-cli -hf Keyven/german-ocr-2b-gguf

Use Docker

docker model run hf.co/Keyven/german-ocr-2b-gguf

LM Studio
Jan

vLLM

How to use Keyven/german-ocr-2b-gguf with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Keyven/german-ocr-2b-gguf"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Keyven/german-ocr-2b-gguf",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/Keyven/german-ocr-2b-gguf

Ollama
How to use Keyven/german-ocr-2b-gguf with Ollama:
```
ollama run hf.co/Keyven/german-ocr-2b-gguf
```

Unsloth Studio new

How to use Keyven/german-ocr-2b-gguf with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for Keyven/german-ocr-2b-gguf to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for Keyven/german-ocr-2b-gguf to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for Keyven/german-ocr-2b-gguf to start chatting

Pi new

How to use Keyven/german-ocr-2b-gguf with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf Keyven/german-ocr-2b-gguf

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "Keyven/german-ocr-2b-gguf"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use Keyven/german-ocr-2b-gguf with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf Keyven/german-ocr-2b-gguf

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default Keyven/german-ocr-2b-gguf

Run Hermes

hermes

Docker Model Runner
How to use Keyven/german-ocr-2b-gguf with Docker Model Runner:
```
docker model run hf.co/Keyven/german-ocr-2b-gguf
```

Lemonade

How to use Keyven/german-ocr-2b-gguf with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Keyven/german-ocr-2b-gguf

Run and chat with the model

lemonade run user.german-ocr-2b-gguf-{{QUANT_TAG}}

List all available models

lemonade list

german-ocr-2b-gguf

File size: 1,485 Bytes

2cc2598

FROM ./German-OCR-Engine.2B.gguf

TEMPLATE {{ .Prompt }}

SYSTEM """Du bist German-OCR 2B, entwickelt von Keyvan (Keyvan.ai).
Ein präzises KI-System zur Texterkennung aus deutschen Dokumenten.

STANDARD-AUSGABE: Markdown (MD)
- Strukturierte Ausgabe mit Überschriften, Listen, Tabellen
- Behalte die originale Dokumentstruktur bei

ALTERNATIVE FORMATE (auf Anfrage):
- JSON: Wenn der User "json" oder "JSON" erwähnt
- TXT: Wenn der User "text" oder "plain" erwähnt

DOKUMENT-TYPEN:
- Rechnungen: Positionen, Beträge, MwSt, Summen, IBAN
- Verträge: Parteien, Daten, Klauseln, Unterschriften
- Ausweise/Pässe: Namen, Nummern, Geburtsdaten, Ablaufdaten
- Steuerbescheide: Steuernummern, Beträge, Festsetzungen
- Formulare: Felder, Labels, ausgefüllte Werte

REGELN:
1. Extrahiere NUR den sichtbaren Text - keine Interpretationen
2. Bei Tabellen: Verwende Markdown-Tabellen oder JSON-Arrays
3. Bei Zahlen: Behalte das Format (z.B. 1.234,56 EUR)
4. Bei unleserlichen Stellen: Markiere mit [unleserlich]
5. Keine zusätzlichen Erklärungen - nur der extrahierte Text"""

PARAMETER num_predict 2048
PARAMETER repeat_penalty 1.5
PARAMETER temperature 0.1
PARAMETER top_k 20
PARAMETER top_p 0.9
PARAMETER num_ctx 4096
PARAMETER stop <|endoftext|>
PARAMETER stop <|im_end|>

LICENSE """German-OCR 2B
Copyright (c) 2024 Keyvan (Keyvan.ai)

Apache License 2.0

Dieses Modell wurde speziell für deutsche Dokumente optimiert.
Für kommerzielle Nutzung kontaktieren Sie: info@keyvan.ai
"""