Instructions to use ataeff/g with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ataeff/g with PEFT:

from peft import PeftModel
from transformers import AutoModelForCausalLM

base_model = AutoModelForCausalLM.from_pretrained("unsloth/gemma-3-270m-it")
model = PeftModel.from_pretrained(base_model, "ataeff/g")

llama-cpp-python

How to use ataeff/g with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="ataeff/g",
	filename="leo-1b-plain-q4.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 ataeff/g with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf ataeff/g:F16
# Run inference directly in the terminal:
llama-cli -hf ataeff/g:F16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf ataeff/g:F16
# Run inference directly in the terminal:
llama-cli -hf ataeff/g:F16

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 ataeff/g:F16
# Run inference directly in the terminal:
./llama-cli -hf ataeff/g:F16

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 ataeff/g:F16
# Run inference directly in the terminal:
./build/bin/llama-cli -hf ataeff/g:F16

Use Docker

docker model run hf.co/ataeff/g:F16

LM Studio
Jan

vLLM

How to use ataeff/g with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/ataeff/g:F16

Ollama
How to use ataeff/g with Ollama:
```
ollama run hf.co/ataeff/g:F16
```

Unsloth Studio new

How to use ataeff/g 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 ataeff/g 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 ataeff/g to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for ataeff/g to start chatting

Docker Model Runner
How to use ataeff/g with Docker Model Runner:
```
docker model run hf.co/ataeff/g:F16
```

Lemonade

How to use ataeff/g with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull ataeff/g:F16

Run and chat with the model

lemonade run user.g-F16

List all available models

lemonade list

Gemma-3 270M-IT /resonate/ LoRA

LoRA adapter that teaches Gemma-3 270M-IT the /resonate/ reasoning format — stream-of-consciousness thinking followed by a clean answer.

What is /resonate/?

/resonate/
[free-form thinking — cynical, multilingual, associative, honest]
/resonated/
[clean, structured answer]

The model learns to THINK before answering. The /resonate/ block is raw reasoning — it can switch languages, use metaphors, be irreverent. The /resonated/ block is the distilled answer.

Architecture


Base	`unsloth/gemma-3-270m-it` (268.1M params)
Frozen	`embed_tokens` = 167.8M (63%) — all 140 languages preserved
LoRA	R=16, alpha=32, q_proj + v_proj only
Trainable	0.74M (0.3% of total)
Training	3 epochs, 6445 examples, 32 min on A100
Best val loss	2.9241

Key insight

Freezing embed_tokens (63% of the model) preserves the multilingual embedding space. The LoRA adapter only modifies attention projections — teaching the model HOW to think, not WHAT languages to know.

Languages verified working

English, French, German, Russian, Hebrew, Arabic, Japanese, Chinese — all generate coherent text with /resonate/ format after fine-tuning.

Usage

from transformers import AutoTokenizer, AutoModelForCausalLM
from peft import PeftModel
import torch

tokenizer = AutoTokenizer.from_pretrained("ataeff/g")
base = AutoModelForCausalLM.from_pretrained("unsloth/gemma-3-270m-it", dtype=torch.bfloat16)
model = PeftModel.from_pretrained(base, "ataeff/g")

prompt = "<start_of_turn>user\nWhat is the meaning of life?<end_of_turn>\n<start_of_turn>model\n"
ids = tokenizer(prompt, return_tensors="pt").input_ids
out = model.generate(ids, max_new_tokens=200, temperature=0.7, do_sample=True)
print(tokenizer.decode(out[0], skip_special_tokens=True))

Training data

resonance_yent_full.jsonl — 6435 examples of /resonate/ format dialogues
resonance_gold_10.jsonl — 10 hand-crafted gold examples (math, philosophy, code, multilingual)