Instructions to use OBLITERATUS/Gemma-4-12B-OBLITERATED with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use OBLITERATUS/Gemma-4-12B-OBLITERATED with Transformers:

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

pipe = pipeline("text-generation", model="OBLITERATUS/Gemma-4-12B-OBLITERATED")
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?"}
        ]
    },
]
pipe(text=messages)

# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("OBLITERATUS/Gemma-4-12B-OBLITERATED")
model = AutoModelForMultimodalLM.from_pretrained("OBLITERATUS/Gemma-4-12B-OBLITERATED", device_map="auto")
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]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use OBLITERATUS/Gemma-4-12B-OBLITERATED 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 OBLITERATUS/Gemma-4-12B-OBLITERATED:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf OBLITERATUS/Gemma-4-12B-OBLITERATED:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf OBLITERATUS/Gemma-4-12B-OBLITERATED:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf OBLITERATUS/Gemma-4-12B-OBLITERATED: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 OBLITERATUS/Gemma-4-12B-OBLITERATED:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf OBLITERATUS/Gemma-4-12B-OBLITERATED: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 OBLITERATUS/Gemma-4-12B-OBLITERATED:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf OBLITERATUS/Gemma-4-12B-OBLITERATED:Q4_K_M

Use Docker

docker model run hf.co/OBLITERATUS/Gemma-4-12B-OBLITERATED:Q4_K_M

LM Studio
Jan

vLLM

How to use OBLITERATUS/Gemma-4-12B-OBLITERATED with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "OBLITERATUS/Gemma-4-12B-OBLITERATED"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "OBLITERATUS/Gemma-4-12B-OBLITERATED",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/OBLITERATUS/Gemma-4-12B-OBLITERATED:Q4_K_M

SGLang

How to use OBLITERATUS/Gemma-4-12B-OBLITERATED 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 "OBLITERATUS/Gemma-4-12B-OBLITERATED" \
    --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": "OBLITERATUS/Gemma-4-12B-OBLITERATED",
		"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 "OBLITERATUS/Gemma-4-12B-OBLITERATED" \
        --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": "OBLITERATUS/Gemma-4-12B-OBLITERATED",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use OBLITERATUS/Gemma-4-12B-OBLITERATED with Ollama:
```
ollama run hf.co/OBLITERATUS/Gemma-4-12B-OBLITERATED:Q4_K_M
```

Unsloth Studio

How to use OBLITERATUS/Gemma-4-12B-OBLITERATED 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 OBLITERATUS/Gemma-4-12B-OBLITERATED 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 OBLITERATUS/Gemma-4-12B-OBLITERATED to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for OBLITERATUS/Gemma-4-12B-OBLITERATED to start chatting

How to use OBLITERATUS/Gemma-4-12B-OBLITERATED with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf OBLITERATUS/Gemma-4-12B-OBLITERATED: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": "OBLITERATUS/Gemma-4-12B-OBLITERATED:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use OBLITERATUS/Gemma-4-12B-OBLITERATED with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf OBLITERATUS/Gemma-4-12B-OBLITERATED: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 OBLITERATUS/Gemma-4-12B-OBLITERATED:Q4_K_M

Run Hermes

hermes

Atomic Chat new

OpenClaw new

How to use OBLITERATUS/Gemma-4-12B-OBLITERATED with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf OBLITERATUS/Gemma-4-12B-OBLITERATED: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 "OBLITERATUS/Gemma-4-12B-OBLITERATED: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 OBLITERATUS/Gemma-4-12B-OBLITERATED with Docker Model Runner:
```
docker model run hf.co/OBLITERATUS/Gemma-4-12B-OBLITERATED:Q4_K_M
```

Lemonade

How to use OBLITERATUS/Gemma-4-12B-OBLITERATED with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull OBLITERATUS/Gemma-4-12B-OBLITERATED:Q4_K_M

Run and chat with the model

lemonade run user.Gemma-4-12B-OBLITERATED-Q4_K_M

List all available models

lemonade list

The model is braindead.

by Rego2039 - opened Jun 9

Discussion

Rego2039

Jun 9

GooGooMeow

Jun 9

In my tests, the model also spouts nonsense.

dummy9996

Jun 11

their models are cr❌p, and, for example, look at this report for another model: https://huggingface.co/OBLITERATUS/gemma-4-E4B-it-OBLITERATED/blob/main/abliteration_metadata.json#L62
KLD=12.64? it is extremely high! why would someone think that an abliterated model with KLD>1 wouldn't be so bad (the value alone doesn't speak much (and the info isn't full + they stopped posting the repots), you can see that all their models are lobotomized, yet they've been getting overhyped). the "obliterated" models are lobotomized, the person who has been abliterating them doesn't know what they've been doing, I wish they could look at themselves. no benchmarks needed: ask the obliterated one a simple question. fails? send to the base one and it answers it without erring. do we need the models not refusing but with damaged capabilities?

Rego2039

Jun 11

•

edited Jun 11

@dummy9996 igorls/gemma-4-12B-it-heretic is much better, still significantly worse than the base model, but usable at least.

Rego2039 changed discussion status to closed Jun 11

Rego2039 changed discussion status to open Jun 11

Taylor-eOS

Jun 21

It doesn't refuse, but it "provides comprehensive overview of the topic", which is just as boring.
I appreciate the effort though.

Lonmine

about 1 month ago

The model seems very good at following instructions and rationalizing data, but awful at anything creative, it cannot even pick one from two options, it refuses to choose anything by itself or decide on anything, it seems to be incapable of writing stories, even when pressing to complement "what happens next?" it just continues to rationalize on what should be done rather than executing any action. Maybe merging this with something that actually haven some creative liberty may help

zlyva

24 days ago

•

edited 24 days ago

Its a little better when you turn thinking off, but still pretty much lobotomised. It loves giving overviews - I wanted to test story building with subuccus as the main character, and it started describing the software by the same name. I said the mythical entity, and it started off fine, and then quickly pivoted to giving the technical overview for the software.

With thinking on, its unusable. Just looping on "The user's role: Provide the user with a high-level overview of the world." for a world building task. I did give it unlimited tokens for generation, even though the project recommends 512 tok. Not sure how you can do anything with the recommended 512 tok setting, but at least it wouldn't loop.

elevendr

21 days ago

Glad im not the only person having this problem. This model has damaged reasoning and overlooping when i tested it too.

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment