Instructions to use Zhantas/DeepGemma-2B-Reasoning with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Zhantas/DeepGemma-2B-Reasoning with Transformers:

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("Zhantas/DeepGemma-2B-Reasoning", dtype="auto")

llama-cpp-python

How to use Zhantas/DeepGemma-2B-Reasoning with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="Zhantas/DeepGemma-2B-Reasoning",
	filename="gemma4_e2b-q4_k_m.gguf",
)

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

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use Zhantas/DeepGemma-2B-Reasoning with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf Zhantas/DeepGemma-2B-Reasoning:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf Zhantas/DeepGemma-2B-Reasoning:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf Zhantas/DeepGemma-2B-Reasoning:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf Zhantas/DeepGemma-2B-Reasoning: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 Zhantas/DeepGemma-2B-Reasoning:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf Zhantas/DeepGemma-2B-Reasoning: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 Zhantas/DeepGemma-2B-Reasoning:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf Zhantas/DeepGemma-2B-Reasoning:Q4_K_M

Use Docker

docker model run hf.co/Zhantas/DeepGemma-2B-Reasoning:Q4_K_M

LM Studio
Jan
Ollama
How to use Zhantas/DeepGemma-2B-Reasoning with Ollama:
```
ollama run hf.co/Zhantas/DeepGemma-2B-Reasoning:Q4_K_M
```

Unsloth Studio new

How to use Zhantas/DeepGemma-2B-Reasoning 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 Zhantas/DeepGemma-2B-Reasoning 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 Zhantas/DeepGemma-2B-Reasoning to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for Zhantas/DeepGemma-2B-Reasoning to start chatting

Pi new

How to use Zhantas/DeepGemma-2B-Reasoning with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf Zhantas/DeepGemma-2B-Reasoning: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": "Zhantas/DeepGemma-2B-Reasoning:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use Zhantas/DeepGemma-2B-Reasoning with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf Zhantas/DeepGemma-2B-Reasoning: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 Zhantas/DeepGemma-2B-Reasoning:Q4_K_M

Run Hermes

hermes

Docker Model Runner
How to use Zhantas/DeepGemma-2B-Reasoning with Docker Model Runner:
```
docker model run hf.co/Zhantas/DeepGemma-2B-Reasoning:Q4_K_M
```

Lemonade

How to use Zhantas/DeepGemma-2B-Reasoning with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Zhantas/DeepGemma-2B-Reasoning:Q4_K_M

Run and chat with the model

lemonade run user.DeepGemma-2B-Reasoning-Q4_K_M

List all available models

lemonade list

🧠 DeepGemma-2B-Reasoning

DeepGemma-2B-Reasoning is a deeply fine-tuned version of google/gemma-4-E2B-it, optimized for step-by-step reasoning (Chain-of-Thought). Trained via knowledge distillation on datasets generated by Claude Opus, Qwen3.5, and KIMI.

The model generates internal reasoning inside <thought> / <think> tags before answering, significantly improving logical and mathematical response quality.

🏆 Benchmarks (GSM8K)

Model	GSM8K (Accuracy)	Improvement
`google/gemma-4-E2B-it` (Base)	30.0%	-
DeepGemma-2B-Reasoning (Ours)	44.0%	+14.0% 🚀

🛠 Training Details

Training was conducted using Unsloth (QLoRA) on an RTX 4090 (24GB VRAM).

Method: QLoRA (4-bit quantization, BF16 adapters)
LoRA Parameters: Rank = 48, Alpha = 48
Epochs: 2 | Global Steps: 4672 | Learning Rate: 2e-4
Final Loss: 1.24

🗜️ GGUF Version (llama.cpp)

A quantized Q4_K_M GGUF version is available directly in this repo.

File: gemma4_e2b-q4_k_m.gguf (~4.7 GB) Quantization: llama.cpp Q4_K_M Merge: Full LoRA merge before quantization (Unsloth)

⚡ Performance (RTX 4090, llama.cpp, ngl=999)

Metric	Speed
Prompt processing	~400 tok/s
Generation	~239–262 tok/s
Context	4096 tokens

Usage (llama.cpp)

./llama-cli -m gemma4_e2b-q4_k_m.gguf \
  -p "<start_of_turn>user\nYour question here<end_of_turn>\n<start_of_turn>model\n" \
  -n 512 -ngl 999 -c 4096

💻 Usage (Transformers / Unsloth)

from unsloth import FastVisionModel

model, tokenizer = FastVisionModel.from_pretrained(
    model_name="Zhantas/DeepGemma-2B-Reasoning",
    max_seq_length=2048,
    load_in_4bit=True,
)
FastVisionModel.for_inference(model)

question = "I had 3 apples. I ate one, and then bought as many as I had left. How many apples do I have now? Reason step by step."
prompt = f"<start_of_turn>user\n{question}<end_of_turn>\n<start_of_turn>model\n"

inputs = tokenizer(text=prompt, return_tensors="pt").to("cuda")
outputs = model.generate(**inputs, max_new_tokens=1024, temperature=0.3, repetition_penalty=1.1)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[1]:], skip_special_tokens=True))

⚠️ Limitations

Prone to "overthinking" on simple tasks. Best suited for logic puzzles, coding, and mathematics.

Downloads last month: 52

GGUF

Model size

5B params

Architecture

gemma4

Hardware compatibility

4-bit

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for Zhantas/DeepGemma-2B-Reasoning

Base model

google/gemma-4-E2B

Finetuned

google/gemma-4-E2B-it

Quantized

(196)

this model

Quantizations

1 model

Zhantas
/

DeepGemma-2B-Reasoning