Instructions to use L-Alchemyst/Coder3101-Gemma-4-E2B-it-heretic-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use L-Alchemyst/Coder3101-Gemma-4-E2B-it-heretic-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 L-Alchemyst/Coder3101-Gemma-4-E2B-it-heretic-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf L-Alchemyst/Coder3101-Gemma-4-E2B-it-heretic-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 L-Alchemyst/Coder3101-Gemma-4-E2B-it-heretic-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf L-Alchemyst/Coder3101-Gemma-4-E2B-it-heretic-GGUF: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 L-Alchemyst/Coder3101-Gemma-4-E2B-it-heretic-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama cli -hf L-Alchemyst/Coder3101-Gemma-4-E2B-it-heretic-GGUF:Q4_K_M

Use Docker

docker model run hf.co/L-Alchemyst/Coder3101-Gemma-4-E2B-it-heretic-GGUF:Q4_K_M

Llama new
LM Studio
Jan
Ollama
How to use L-Alchemyst/Coder3101-Gemma-4-E2B-it-heretic-GGUF with Ollama:
```
ollama run hf.co/L-Alchemyst/Coder3101-Gemma-4-E2B-it-heretic-GGUF:Q4_K_M
```
Unsloth Desktop

How to use L-Alchemyst/Coder3101-Gemma-4-E2B-it-heretic-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf L-Alchemyst/Coder3101-Gemma-4-E2B-it-heretic-GGUF: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": "L-Alchemyst/Coder3101-Gemma-4-E2B-it-heretic-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use L-Alchemyst/Coder3101-Gemma-4-E2B-it-heretic-GGUF with Docker Model Runner:
```
docker model run hf.co/L-Alchemyst/Coder3101-Gemma-4-E2B-it-heretic-GGUF:Q4_K_M
```

Lemonade

How to use L-Alchemyst/Coder3101-Gemma-4-E2B-it-heretic-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull L-Alchemyst/Coder3101-Gemma-4-E2B-it-heretic-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.Coder3101-Gemma-4-E2B-it-heretic-GGUF-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use L-Alchemyst/Coder3101-Gemma-4-E2B-it-heretic-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 L-Alchemyst/Coder3101-Gemma-4-E2B-it-heretic-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 L-Alchemyst/Coder3101-Gemma-4-E2B-it-heretic-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use L-Alchemyst/Coder3101-Gemma-4-E2B-it-heretic-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf L-Alchemyst/Coder3101-Gemma-4-E2B-it-heretic-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 "L-Alchemyst/Coder3101-Gemma-4-E2B-it-heretic-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"

gemma-4-E2B-it-heretic GGUF

Quantized GGUF versions of coder3101/gemma-4-E2B-it-heretic, a Heretic ARA-ablated version of Google's Gemma 4 E2B-it.

Quantized by: L-Alchemyst
Original model: coder3101/gemma-4-E2B-it-heretic

Available Quants

Quant Type	File	Size	Use Case
Q4_0-PURE	`gemma-4-E2B-it-heretic-Q4_0-PURE.gguf`	~2.65 GB	Adreno 830 / mobile inference — uniform block layout optimized for Vulkan/mobile GPU kernels
Q4_K_M	`gemma-4-E2B-it-heretic-Q4_K_M.gguf`	~3.43 GB	Best quality/size tradeoff for desktop CPU
Q5_K_M	`gemma-4-E2B-it-heretic-Q5_K_M.gguf`	~3.63 GB	Higher quality, still RAM-friendly
Q6_K	`gemma-4-E2B-it-heretic-Q6_K.gguf`	~3.85 GB	Near-lossless, good for GPU offload
Q8_0	`gemma-4-E2B-it-heretic-Q8_0.gguf`	~4.97 GB	Effectively lossless, reference quality
FP16	`gemma-4-E2B-it-heretic-FP16.gguf`	~9.31 GB	Full precision

Mobile / Adreno 830 Users

The Q4_0-PURE quant is specifically recommended for Snapdragon 8 Elite devices (Adreno 830). Its uniform 4-bit block layout maps directly to mobile Vulkan shader paths in llama.cpp, giving the best tokens/sec on-device vs. the K-quant variants.

Estimated runtime RAM: ~3–4.5 GB including KV cache — fits comfortably on 8 GB+ devices.

Compatible with: llama.cpp, MLC LLM probably, Termux + llama.cpp

Achieves prefill numbers of 328.35 tps and generation speeds between 13-17 tps running on a Samsung S25 Edge with these arguments:

LD_LIBRARY_PATH=/vendor/lib64:$PREFIX/lib
./build-android/bin/llama-server
-m ~/models/Coder3101-Gemma-4-E2B-it-heretic-Q4_0-PURE_1.gguf
--alias Gemma-4-E2B-Heretic
-ngl 99
-b 1024 -ub 256
-c 8192
--ctx-checkpoints 2
-np 1
--temp 1.0
--min-p 0.05
-cram 600
--repeat-penalty 1.15 --host 0.0.0.0 --port 8080

About the Original Model

Gemma 4 E2B-it is Google DeepMind's smallest Gemma 4 model (~2B effective parameters via Per-Layer Embeddings). The Heretic variant applies ARA (Arbitrary-Rank Ablation) to bypass safety restrictions.

How to Run

llama-cli -m gemma-4-E2B-it-heretic-Q4_0-PURE.gguf \
  -n 512 \
  -p "Your prompt here"

Downloads last month: 358

GGUF

Model size

5B params

Architecture

gemma4

Hardware compatibility

4-bit

5-bit

6-bit

8-bit

16-bit

Inference Providers NEW

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

Model tree for L-Alchemyst/Coder3101-Gemma-4-E2B-it-heretic-GGUF

Base model

google/gemma-4-E2B

Finetuned

google/gemma-4-E2B-it

Finetuned

coder3101/gemma-4-E2B-it-heretic

Quantized

(6)

this model