Instructions to use GermannM/Kenga-Coder-7B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries
PEFT
How to use GermannM/Kenga-Coder-7B with PEFT:
```
Task type is invalid.
```
Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use GermannM/Kenga-Coder-7B with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf GermannM/Kenga-Coder-7B:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf GermannM/Kenga-Coder-7B:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf GermannM/Kenga-Coder-7B:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf GermannM/Kenga-Coder-7B:Q4_K_M

Build from source code

git clone https://github.com/ggml-org/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-app
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama serve -hf GermannM/Kenga-Coder-7B:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama cli -hf GermannM/Kenga-Coder-7B:Q4_K_M

Use Docker

docker model run hf.co/GermannM/Kenga-Coder-7B:Q4_K_M

Llama new
LM Studio
Jan
Ollama
How to use GermannM/Kenga-Coder-7B with Ollama:
```
ollama run hf.co/GermannM/Kenga-Coder-7B:Q4_K_M
```
Unsloth Desktop

How to use GermannM/Kenga-Coder-7B with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf GermannM/Kenga-Coder-7B:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "GermannM/Kenga-Coder-7B:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use GermannM/Kenga-Coder-7B with Docker Model Runner:
```
docker model run hf.co/GermannM/Kenga-Coder-7B:Q4_K_M
```

Lemonade

How to use GermannM/Kenga-Coder-7B with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull GermannM/Kenga-Coder-7B:Q4_K_M

Run and chat with the model

lemonade run user.Kenga-Coder-7B-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use GermannM/Kenga-Coder-7B with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf GermannM/Kenga-Coder-7B:Q4_K_M

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 GermannM/Kenga-Coder-7B:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use GermannM/Kenga-Coder-7B with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf GermannM/Kenga-Coder-7B:Q4_K_M

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "GermannM/Kenga-Coder-7B:Q4_K_M" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

Kenga-Coder-7B

File size: 2,091 Bytes

---
license: apache-2.0
language:
- ru
- en
base_model: Qwen/Qwen2.5-Coder-7B-Instruct
library_name: peft
tags:
- kenga
- kenga-lang
- russian
- z-system
- code
- qwen2.5-coder
- gguf
- conversational
---

# Kenga-Coder-7B

Кодер-модель Кенги на базе **Qwen2.5-Coder-7B-Instruct**, дообученная на
верифицированном датасете (язык Kenga + Python-ката + Z-система + идентичность).

## Что внутри

- **~640 программ на языке Kenga, проверенных компилятором** (`kenga-lite`):
  каждая реально скомпилирована и запущена (compiler-in-the-loop).
- **700 Python-задач**, сгенерированных и проверенных исполнением.
- Документация языка Kenga (struct, enum, map, events, stdlib math/list/agent, tensor).
- Z-система и материалы kenga-lang, идентичность Кенги, Россия/культура, математика.
- Принцип M5.3: правильность — не мнение, а факт (программа запускается, ответ проверен).

## Обучение

- База: `Qwen/Qwen2.5-Coder-7B-Instruct`
- Метод: QLoRA (r=32, alpha=64, NF4 double-quant), bf16, grad-checkpointing
- Железо: RTX 4070 SUPER 12GB, локально
- Датасет: 6394 диалога, 3 эпохи, seq 2048
- Квантизация: **imatrix** (Q4_K_M / Q5_K_M)

## Файлы

- `kenga-coder-7b-q4_k_m.gguf` — ~4.4GB, для 6GB GPU
- `kenga-coder-7b-q5_k_m.gguf` — выше качество, ~5.3GB
- `adapter/` — LoRA-адаптер

## Запуск

```bash
ollama run hf.co/GermannM/Kenga-Coder-7B:Q4_K_M
```

```python
from transformers import AutoModelForCausalLM, AutoTokenizer
tok = AutoTokenizer.from_pretrained("GermannM/Kenga-Coder-7B")
model = AutoModelForCausalLM.from_pretrained("GermannM/Kenga-Coder-7B")
```

---

Создатель проекта: Герман Янтарас (GermannM)