Instructions to use kirilldual0987/rumeni-1-nano-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use kirilldual0987/rumeni-1-nano-GGUF with Transformers:

# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("kirilldual0987/rumeni-1-nano-GGUF")
model = AutoModelForMultimodalLM.from_pretrained("kirilldual0987/rumeni-1-nano-GGUF")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
inputs = processor.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

llama-cpp-python

How to use kirilldual0987/rumeni-1-nano-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="kirilldual0987/rumeni-1-nano-GGUF",
	filename="rumeni-1-nano-F16.gguf",
)

llm.create_chat_completion(
	messages = "No input example has been defined for this model task."
)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use kirilldual0987/rumeni-1-nano-GGUF 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 kirilldual0987/rumeni-1-nano-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf kirilldual0987/rumeni-1-nano-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf kirilldual0987/rumeni-1-nano-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf kirilldual0987/rumeni-1-nano-GGUF:Q4_K_M

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 kirilldual0987/rumeni-1-nano-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf kirilldual0987/rumeni-1-nano-GGUF:Q4_K_M

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 kirilldual0987/rumeni-1-nano-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf kirilldual0987/rumeni-1-nano-GGUF:Q4_K_M

Use Docker

docker model run hf.co/kirilldual0987/rumeni-1-nano-GGUF:Q4_K_M

LM Studio
Jan
Ollama
How to use kirilldual0987/rumeni-1-nano-GGUF with Ollama:
```
ollama run hf.co/kirilldual0987/rumeni-1-nano-GGUF:Q4_K_M
```

Unsloth Studio

How to use kirilldual0987/rumeni-1-nano-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 kirilldual0987/rumeni-1-nano-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 kirilldual0987/rumeni-1-nano-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for kirilldual0987/rumeni-1-nano-GGUF to start chatting

How to use kirilldual0987/rumeni-1-nano-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf kirilldual0987/rumeni-1-nano-GGUF:Q4_K_M

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": "kirilldual0987/rumeni-1-nano-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use kirilldual0987/rumeni-1-nano-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf kirilldual0987/rumeni-1-nano-GGUF: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 kirilldual0987/rumeni-1-nano-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat new

OpenClaw new

How to use kirilldual0987/rumeni-1-nano-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf kirilldual0987/rumeni-1-nano-GGUF: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 "kirilldual0987/rumeni-1-nano-GGUF: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"

Docker Model Runner
How to use kirilldual0987/rumeni-1-nano-GGUF with Docker Model Runner:
```
docker model run hf.co/kirilldual0987/rumeni-1-nano-GGUF:Q4_K_M
```

Lemonade

How to use kirilldual0987/rumeni-1-nano-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull kirilldual0987/rumeni-1-nano-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.rumeni-1-nano-GGUF-Q4_K_M

List all available models

lemonade list

Rumeni 1 nano

🚀 Technical Highlights

Frankenstein Merge Architecture - Modular component integration
Memory-Efficient - Runs on 16GB RAM with CPU inference
GGUF Compatible - Can be quantized to Q5_K_M (~4-6GB)
Extensible - Easy to add new LoRA adapters
Privacy-First - All processing happens locally

Component Breakdown

Component	Weight	Source	Purpose
`vision_tower`	100%	Google Gemma 4 E2B	Image processing
`audio_tower`	100%	Google Gemma 4 E2B	Audio processing
`embed_tokens`	100%	Google Gemma 4 E2B	Token embeddings
`language_model.layers`	60% Rumeni + 40% Opus	HuiHui + Claude Opus	Text generation + reasoning
`language_model.norm`	70% Rumeni + 30% Opus	HuiHui + Claude Opus	Layer normalization
`lm_head`	50% Rumeni + 50% Opus	HuiHui + Claude Opus	Output projection

1. Feature Comparison

Feature	Google Gemma 4 E2B	HuiHui Abliterated	Opus Reasoning	Rumeni	GPT-4o	Gemini Pro
Vision (Images)	✅	❌	❌	✅	✅	✅
Audio Processing	✅	❌	❌	✅	✅	✅
Uncensored Text	❌	✅		✅	❌	❌
Advanced Reasoning	⭐⭐⭐	⭐⭐	⭐⭐⭐⭐⭐	⭐⭐⭐⭐	⭐⭐⭐⭐⭐	⭐⭐⭐⭐⭐
Russian Language	⭐⭐	⭐⭐⭐	⭐⭐⭐	⭐⭐⭐⭐	⭐⭐⭐⭐	⭐⭐⭐⭐
Local Deployment	✅	✅	✅	✅	❌	❌
Privacy	✅	✅	✅	✅		❌
Free & Open Source	✅	✅	✅	✅	❌	❌
Model Size	~10 GB	~10 GB	~10 GB	~10 GB	N/A	N/A
Hardware Requirements	16GB RAM	16GB RAM	16GB RAM	16GB RAM	Cloud	Cloud

2. Performance Benchmarks

Tested on Intel N100 (Beelink mini PC, 16GB RAM, CPU-only inference)

Task	Google Gemma 4	HuiHui	Opus	Rumeni	GPT-4o (API)
Text Generation (512 tokens)	42s	45s	48s	47s	~2s
Image Analysis	58s	N/A	N/A	61s	~3s
Audio Transcription (1 min)	85s	N/A	N/A	92s	~5s
Math Problem (Complex)	28s	31s	22s	25s	~1s
Code Generation	38s	41s	35s	40s	~2s
Reasoning (Logic)	⭐⭐⭐	⭐⭐⭐	⭐⭐⭐⭐⭐	⭐⭐⭐⭐	⭐⭐⭐⭐
RAM Usage	12 GB	12 GB	13 GB	14 GB	N/A
Disk Space	10 GB	10 GB	10 GB	10 GB	N/A

3. Capability Comparison

Capability	Google Gemma 4 E2B	HuiHui Abliterated	Opus Reasoning	Rumeni	LLaVA-1.6	Claude 3.5 Sonnet
Multimodal Input	✅ Vision + Audio	Text only	❌ Text only	✅ Vision + Audio	✅ Vision only	✅ Vision only
Content Restrictions	High	None	High	None	Medium	High
Step-by-Step Reasoning	Basic	Basic	Advanced	Advanced	Basic	Advanced
Creative Writing	⭐⭐⭐⭐	⭐⭐⭐⭐⭐	⭐⭐⭐	⭐⭐⭐⭐⭐	⭐⭐⭐	⭐⭐⭐⭐⭐
Mathematical Accuracy	72%	68%	89%	85%	65%	92%
Code Generation	⭐⭐⭐⭐	⭐⭐⭐⭐	⭐⭐⭐⭐	⭐⭐⭐⭐	⭐⭐⭐	⭐⭐⭐⭐⭐
Russian Proficiency	⭐⭐⭐	⭐⭐⭐	⭐⭐⭐	⭐⭐⭐⭐	⭐⭐	⭐⭐⭐
Offline Operation	✅	✅	✅	✅	✅	❌
Customizable	✅	✅	✅	✅	✅	❌

4. Component Breakdown

Component	Weight	Source	Purpose
`vision_tower`	100%	Google Gemma 4 E2B	Image processing
`audio_tower`	100%	Google Gemma 4 E2B	Audio processing
`embed_tokens`	100%	Google Gemma 4 E2B	Token embeddings
`language_model.layers`	60% Rumeni + 40% Opus	HuiHui + Claude Opus	Text generation + reasoning
`language_model.norm`	70% Rumeni + 30% Opus	HuiHui + Claude Opus	Layer normalization
`lm_head`	50% Rumeni + 50% Opus	HuiHui + Claude Opus	Output projection

5. Use Case Comparison

Use Case	Google Gemma 4	HuiHui	Opus	Rumeni	Best For
Personal Assistant	⭐⭐⭐	⭐⭐⭐⭐	⭐⭐⭐⭐	⭐⭐⭐⭐	Rumeni - all-in-one
Content Creation	⭐⭐⭐	⭐⭐⭐⭐⭐	⭐⭐⭐⭐	⭐⭐⭐⭐⭐	HuiHui/Rumeni - uncensored
Math & Logic	⭐⭐⭐	⭐⭐⭐	⭐⭐⭐⭐⭐	⭐⭐⭐⭐	Opus/Rumeni - reasoning
Image Analysis	⭐⭐⭐⭐⭐	❌	❌	⭐⭐⭐⭐	Google/Rumeni - vision
Audio Transcription	⭐⭐⭐⭐⭐	❌	❌	⭐⭐⭐⭐⭐	Google/Rumeni - audio
Code Generation	⭐⭐⭐⭐	⭐⭐⭐⭐	⭐⭐⭐⭐⭐	⭐⭐⭐⭐	Opus/Rumeni - code