Instructions to use 0xA50C1A1/MN-12B-Nymphaea-RP with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use 0xA50C1A1/MN-12B-Nymphaea-RP with Transformers:

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

pipe = pipeline("text-generation", model="0xA50C1A1/MN-12B-Nymphaea-RP")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("0xA50C1A1/MN-12B-Nymphaea-RP")
model = AutoModelForCausalLM.from_pretrained("0xA50C1A1/MN-12B-Nymphaea-RP")
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]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use 0xA50C1A1/MN-12B-Nymphaea-RP with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "0xA50C1A1/MN-12B-Nymphaea-RP"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "0xA50C1A1/MN-12B-Nymphaea-RP",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/0xA50C1A1/MN-12B-Nymphaea-RP

SGLang

How to use 0xA50C1A1/MN-12B-Nymphaea-RP 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 "0xA50C1A1/MN-12B-Nymphaea-RP" \
    --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": "0xA50C1A1/MN-12B-Nymphaea-RP",
		"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 "0xA50C1A1/MN-12B-Nymphaea-RP" \
        --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": "0xA50C1A1/MN-12B-Nymphaea-RP",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use 0xA50C1A1/MN-12B-Nymphaea-RP with Docker Model Runner:
```
docker model run hf.co/0xA50C1A1/MN-12B-Nymphaea-RP
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

MN-12B-Nymphaea-RP

A fine-tune of Mistral Nemo Instruct 2407 for roleplay and creative writing.

I've trained this mostly for merging with Mistral Nemo fine-tunes.

The SillyTavern preset is available here. For custom presets, please use the Mistral V3-Tekken instruct template.

Training Notes

Trained on the latest iteration of my Darkmere dataset. This version features expanded genre variety, built upon a mix of manually curated synthetics and human-written stories.

The base weights are abliterated via Heretic prior to fine-tuning, so this fine-tune is quite uncensored.

Training Specs

Method:

Training Method: DoRA (Weight-Decomposed LoRA)
Target Modules all-linear
LoRA Rank: 64
LoRA Alpha: 64
LoRA Dropout: 0.05

Hyperparameters:

Batch Size: 2 (Per-device)
Gradient Accumulation: 2
Epochs: 2
Learning Rate: 1e-4
Optimizer: adamw_torch_fused
LR Scheduler: cosine
Noise Level: neftune_noise_alpha=5

Special Thanks

This fine-tune wouldn't be possible without the incredible work of the community:

p-e-w for developing Heretic - an essential tool for censorship removal.
SicariusSicariiStuff for developing SLOP_Detector script.
Mistral AI for their Mistral Nemo weights.
AMD for their Instinct™ MI300X GPU.