Instructions to use bunnycore/Qwen3-4B-Mixture with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use bunnycore/Qwen3-4B-Mixture with Transformers:

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

pipe = pipeline("text-generation", model="bunnycore/Qwen3-4B-Mixture")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("bunnycore/Qwen3-4B-Mixture")
model = AutoModelForCausalLM.from_pretrained("bunnycore/Qwen3-4B-Mixture", device_map="auto")
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]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use bunnycore/Qwen3-4B-Mixture with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "bunnycore/Qwen3-4B-Mixture"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "bunnycore/Qwen3-4B-Mixture",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/bunnycore/Qwen3-4B-Mixture

SGLang

How to use bunnycore/Qwen3-4B-Mixture 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 "bunnycore/Qwen3-4B-Mixture" \
    --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": "bunnycore/Qwen3-4B-Mixture",
		"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 "bunnycore/Qwen3-4B-Mixture" \
        --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": "bunnycore/Qwen3-4B-Mixture",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use bunnycore/Qwen3-4B-Mixture with Docker Model Runner:
```
docker model run hf.co/bunnycore/Qwen3-4B-Mixture
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Qwen3-4B-Mixture is a merged language model built upon the Qwen3-4B architecture. This model is a result of combining several fine-tuned versions of Qwen3-4B, with a particular emphasis on incorporating "uncensored" or less restricted variants. The intention behind this merger is to potentially enhance the model's breadth of knowledge and reduce certain inherent biases or limitations often found in more heavily moderated models.

Merge Method

This model was merged using the Linear merge method.

Models Merged

The following models were included in the merge:

Configuration

The following YAML configuration was used to produce this model:


models:
  - model: mlabonne/Qwen3-4B-abliterated
    parameters:
      weight: 0.4
  - model: ValiantLabs/Qwen3-4B-Esper3
    parameters:
      weight: 0.5
  - model: fakezeta/amoral-Qwen3-4B
    parameters:
      weight: 0.6
  - model: Goekdeniz-Guelmez/Josiefied-Qwen3-4B-abliterated-v1
    parameters:
      weight: 0.4
merge_method: linear
normalize: false
int8_mask: true
dtype: bfloat16