Instructions to use Columbidae/CavesOfQwen3-exp16-Unhealed with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Columbidae/CavesOfQwen3-exp16-Unhealed with Transformers:

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

pipe = pipeline("text-generation", model="Columbidae/CavesOfQwen3-exp16-Unhealed")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Columbidae/CavesOfQwen3-exp16-Unhealed")
model = AutoModelForCausalLM.from_pretrained("Columbidae/CavesOfQwen3-exp16-Unhealed")
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

vLLM

How to use Columbidae/CavesOfQwen3-exp16-Unhealed with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Columbidae/CavesOfQwen3-exp16-Unhealed"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Columbidae/CavesOfQwen3-exp16-Unhealed",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Columbidae/CavesOfQwen3-exp16-Unhealed

SGLang

How to use Columbidae/CavesOfQwen3-exp16-Unhealed 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 "Columbidae/CavesOfQwen3-exp16-Unhealed" \
    --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": "Columbidae/CavesOfQwen3-exp16-Unhealed",
		"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 "Columbidae/CavesOfQwen3-exp16-Unhealed" \
        --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": "Columbidae/CavesOfQwen3-exp16-Unhealed",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Columbidae/CavesOfQwen3-exp16-Unhealed with Docker Model Runner:
```
docker model run hf.co/Columbidae/CavesOfQwen3-exp16-Unhealed
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

This is a copy of KaraKaraWitch/CavesOfQwen3 with the number of active experts turned up to 16. I also transplanted the embeddings from the instruct version of 30B-A3B so the chatml tokens are properly utilized.

This will actually make performance worse if you just use it out of the box (don't let viral reddit posts fool you, turning up experts is one of the first things a lot of us tried to try to stabilize this model). The goal is to try training with more experts active so the model learns to use them properly, hence prepping a version with that already in the config for easy loading.

Downloads last month: 1

Safetensors

Model size

31B params

Tensor type

BF16

Model tree for Columbidae/CavesOfQwen3-exp16-Unhealed

Qwen/Qwen3-30B-A3B

Qwen/Qwen3-30B-A3B-Base

Merge model

this model

Adapters

1 model

Quantizations

3 models