Instructions to use HattoriHanzo1/Liquid-A1B-8B-PL-Coder with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use HattoriHanzo1/Liquid-A1B-8B-PL-Coder with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="HattoriHanzo1/Liquid-A1B-8B-PL-Coder")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("HattoriHanzo1/Liquid-A1B-8B-PL-Coder")
model = AutoModelForCausalLM.from_pretrained("HattoriHanzo1/Liquid-A1B-8B-PL-Coder")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.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(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

llama-cpp-python

How to use HattoriHanzo1/Liquid-A1B-8B-PL-Coder with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="HattoriHanzo1/Liquid-A1B-8B-PL-Coder",
	filename="GGUF/Liquid-A1B-8B-PL-Coder-F16.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use HattoriHanzo1/Liquid-A1B-8B-PL-Coder with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf HattoriHanzo1/Liquid-A1B-8B-PL-Coder:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf HattoriHanzo1/Liquid-A1B-8B-PL-Coder:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf HattoriHanzo1/Liquid-A1B-8B-PL-Coder:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf HattoriHanzo1/Liquid-A1B-8B-PL-Coder: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 HattoriHanzo1/Liquid-A1B-8B-PL-Coder:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf HattoriHanzo1/Liquid-A1B-8B-PL-Coder: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 HattoriHanzo1/Liquid-A1B-8B-PL-Coder:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf HattoriHanzo1/Liquid-A1B-8B-PL-Coder:Q4_K_M

Use Docker

docker model run hf.co/HattoriHanzo1/Liquid-A1B-8B-PL-Coder:Q4_K_M

LM Studio
Jan

vLLM

How to use HattoriHanzo1/Liquid-A1B-8B-PL-Coder with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "HattoriHanzo1/Liquid-A1B-8B-PL-Coder"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "HattoriHanzo1/Liquid-A1B-8B-PL-Coder",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/HattoriHanzo1/Liquid-A1B-8B-PL-Coder:Q4_K_M

SGLang

How to use HattoriHanzo1/Liquid-A1B-8B-PL-Coder with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "HattoriHanzo1/Liquid-A1B-8B-PL-Coder" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "HattoriHanzo1/Liquid-A1B-8B-PL-Coder",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "HattoriHanzo1/Liquid-A1B-8B-PL-Coder" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "HattoriHanzo1/Liquid-A1B-8B-PL-Coder",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use HattoriHanzo1/Liquid-A1B-8B-PL-Coder with Ollama:
```
ollama run hf.co/HattoriHanzo1/Liquid-A1B-8B-PL-Coder:Q4_K_M
```

Unsloth Studio new

How to use HattoriHanzo1/Liquid-A1B-8B-PL-Coder 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 HattoriHanzo1/Liquid-A1B-8B-PL-Coder 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 HattoriHanzo1/Liquid-A1B-8B-PL-Coder to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for HattoriHanzo1/Liquid-A1B-8B-PL-Coder to start chatting

Pi new

How to use HattoriHanzo1/Liquid-A1B-8B-PL-Coder with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf HattoriHanzo1/Liquid-A1B-8B-PL-Coder: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": "HattoriHanzo1/Liquid-A1B-8B-PL-Coder:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use HattoriHanzo1/Liquid-A1B-8B-PL-Coder with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf HattoriHanzo1/Liquid-A1B-8B-PL-Coder: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 HattoriHanzo1/Liquid-A1B-8B-PL-Coder:Q4_K_M

Run Hermes

hermes

Docker Model Runner
How to use HattoriHanzo1/Liquid-A1B-8B-PL-Coder with Docker Model Runner:
```
docker model run hf.co/HattoriHanzo1/Liquid-A1B-8B-PL-Coder:Q4_K_M
```

Lemonade

How to use HattoriHanzo1/Liquid-A1B-8B-PL-Coder with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull HattoriHanzo1/Liquid-A1B-8B-PL-Coder:Q4_K_M

Run and chat with the model

lemonade run user.Liquid-A1B-8B-PL-Coder-Q4_K_M

List all available models

lemonade list

⚔️ Liquid-A1B-8B-PL-Coder (Hattori Hanzo Edition)

📥 Szybkie Pobieranie (GGUF)

Wersje zoptymalizowane do działania na domowym sprzęcie (Llama.cpp, LM Studio, Ollama). Pliki znajdują się w folderze /GGUF:

Model	Rozmiar	Opis	Link
Q6_K	~6.7 GB	Zalecany: Najwyższa precyzja, idealny do kodu	Pobierz Q6
Q4_K_M	~4.8 GB	Szybka inferencja, niski pobór VRAM	Pobierz Q4
F16	~16.7 GB	Master Copy GGUF (Bezstratny)	Pobierz F16

[English Version] | [Polska Wersja Poniżej]

🇺🇸 English Description

Liquid-A1B-8B-PL-Coder is a high-performance adaptation of the Liquid LFM 2.5 8B architecture. Meticulously fine-tuned by HattoriHanzo1, this model combines advanced MoE (Mixture of Experts) logic with deep Polish language understanding.

It is specifically engineered for:

Advanced Programming: Focused on C# (thread-safety, circular buffers, async patterns).
SDR & Electronics: Designed to assist with Software Defined Radio projects and EMS/Arduino development.
Pattern Recognition: Capable of implementing complex logic like binary sequence detection (e.g., "QAPLA" pattern matching).

📊 Training Progress (Loss Matrix)

The model reached an exceptional $Loss$ level, showcasing its superior reasoning capabilities:

Step	Training Loss	Status
600	1.3995	Initialization
880	0.6990	👑 Sweet Spot (Best Logic)
1474	0.7719	Final Tempering (100% Epoch)

🇵🇱 Polska Wersja

Liquid-A1B-8B-PL-Coder to potężna fuzja architektury Liquid AI z polskim zacięciem inżynierskim. Model został wykuty przez HattoriHanzo1, aby służyć jako cyfrowe ostrze w rękach programistów i elektroników.

⚔️ Filozofia "Hattori Hanzo"

Ten model nie powstał w drodze kompromisów. W moim języku nie istnieje słowo "nie da się". Jest tylko "trudniej wykonalne i zajmie więcej czasu". Każdy parametr został dobrany tak, aby model nie tylko "pisał kod", ale rozumiał logikę stojącą za wydajnym przetwarzaniem sygnałów i zarządzaniem pamięcią.

📈 Charakterystyka i Zastosowanie

Model wykazuje rekordowo niski błąd treningowy (0.699), co przekłada się na konkretną wiedzę w obszarach:

C# Expert: Od struktur danych po zaawansowaną wielowątkowość.
SDR / Radio: Wsparcie dla amatorów radiofalowców i budowniczych anten.
Automatyka: Projekty oparte na Arduino i systemach wbudowanych.

"Świat człowieka nie jest tak szybki jak twój w krzemie. Staramy się sparsować nasze światy."

👨‍💻 Creator & Credits

User/Creator: HattoriHanzo1
Architecture: Liquid AI (LFM-2.5-8B MoE)
Dataset: Synthia-Coder-v1.5-I-sharegpt (Modified)
Motto: Qapla'! Sukces to podróż, nie cel.

Downloads last month: 134

Safetensors

Model size

8B params

Tensor type

F32

BF16