nsfw

exl3

conversational

4-bit precision

Model card Files Files and versions

xet

Community

Instructions to use Mawdistical/Draconia-Overdrive-32B_EXL3_4.0bpw_H6 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

Transformers

How to use Mawdistical/Draconia-Overdrive-32B_EXL3_4.0bpw_H6 with Transformers:

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

pipe = pipeline("text-generation", model="Mawdistical/Draconia-Overdrive-32B_EXL3_4.0bpw_H6")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForMultimodalLM

tokenizer = AutoTokenizer.from_pretrained("Mawdistical/Draconia-Overdrive-32B_EXL3_4.0bpw_H6")
model = AutoModelForMultimodalLM.from_pretrained("Mawdistical/Draconia-Overdrive-32B_EXL3_4.0bpw_H6")
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 Mawdistical/Draconia-Overdrive-32B_EXL3_4.0bpw_H6 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Mawdistical/Draconia-Overdrive-32B_EXL3_4.0bpw_H6"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Mawdistical/Draconia-Overdrive-32B_EXL3_4.0bpw_H6",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Mawdistical/Draconia-Overdrive-32B_EXL3_4.0bpw_H6

SGLang

How to use Mawdistical/Draconia-Overdrive-32B_EXL3_4.0bpw_H6 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 "Mawdistical/Draconia-Overdrive-32B_EXL3_4.0bpw_H6" \
    --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": "Mawdistical/Draconia-Overdrive-32B_EXL3_4.0bpw_H6",
		"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 "Mawdistical/Draconia-Overdrive-32B_EXL3_4.0bpw_H6" \
        --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": "Mawdistical/Draconia-Overdrive-32B_EXL3_4.0bpw_H6",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Mawdistical/Draconia-Overdrive-32B_EXL3_4.0bpw_H6 with Docker Model Runner:
```
docker model run hf.co/Mawdistical/Draconia-Overdrive-32B_EXL3_4.0bpw_H6
```

Draconia-Overdrive-32B

Explicit Content Warning

Support Mawdistical finetunes here

"A creation of 'chaos aura' that accentuates draconian fervor."

Draconia-Overdrive-32B is an expressive, creative, and roleplay-driven large language model developed for a wide range of contexts. Drawing inspiration from deep chaos, it brings a fervent, untamed spirit mirroring the energy of relentless draconianism.

✧ Quantized Formats

Original Model:
- Draconia-Overdrive-32B

✧ Recommended Settings

Temperature: 1.0-1.1
Min P: 0.02-0.05
Dynamic Temperature (optional):
- Multiplier: 0.75-0.85
- Base: 1.8
- Length: 4

✧ Sample Presets

Temperature: 1.07
Top-P: 0.92
Min-P: 0.035
Mirostat: 2
Repetition Penalty: 1.12
Dynamic Temperature: on (Multiplier: 0.8, Base: 1.8, Length: 4)

✧ Credits

Model Author: @Mawnipulator
Additional Credit: @xtristan
Government Body:

License: MIT

This model was generously made with compute from Shuttleai.com

Downloads last month: 4

Safetensors

Model size

9B params

Tensor type

F16

I16

Model tree for Mawdistical/Draconia-Overdrive-32B_EXL3_4.0bpw_H6

Base model

Mawdistical/Draconia-Overdrive-32B

Quantized

(12)

this model

Collection including Mawdistical/Draconia-Overdrive-32B_EXL3_4.0bpw_H6

Draconia-Overdrive-32B

Collection

11 items • Updated May 27, 2025 • 2