Not-For-All-Audiences

Instructions to use Kaoeiri/MS-MagpantheonselRP-22B-12.995 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Kaoeiri/MS-MagpantheonselRP-22B-12.995 with Transformers:

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

pipe = pipeline("text-generation", model="Kaoeiri/MS-MagpantheonselRP-22B-12.995")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForMultimodalLM

tokenizer = AutoTokenizer.from_pretrained("Kaoeiri/MS-MagpantheonselRP-22B-12.995")
model = AutoModelForMultimodalLM.from_pretrained("Kaoeiri/MS-MagpantheonselRP-22B-12.995")
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 Kaoeiri/MS-MagpantheonselRP-22B-12.995 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Kaoeiri/MS-MagpantheonselRP-22B-12.995"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Kaoeiri/MS-MagpantheonselRP-22B-12.995",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Kaoeiri/MS-MagpantheonselRP-22B-12.995

SGLang

How to use Kaoeiri/MS-MagpantheonselRP-22B-12.995 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 "Kaoeiri/MS-MagpantheonselRP-22B-12.995" \
    --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": "Kaoeiri/MS-MagpantheonselRP-22B-12.995",
		"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 "Kaoeiri/MS-MagpantheonselRP-22B-12.995" \
        --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": "Kaoeiri/MS-MagpantheonselRP-22B-12.995",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Kaoeiri/MS-MagpantheonselRP-22B-12.995 with Docker Model Runner:
```
docker model run hf.co/Kaoeiri/MS-MagpantheonselRP-22B-12.995
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

merge

This is a merge of pre-trained language models created using mergekit.

Merge Details

Merge Method

This model was merged using the DARE TIES merge method using unsloth/Mistral-Small-Instruct-2409 as a base.

Models Merged

The following models were included in the merge:

Configuration

The following YAML configuration was used to produce this model:

models:
  # Primary Fiction Core (Rebalanced for character focus)
  - model: Kaoeiri/MS_Moingooistral-2409-22B  # Monster fiction core
    parameters:
      weight: 0.22  # Reduced from 0.28 
      density: 0.72 # Adjusted for balance

  - model: Kaoeiri/MS-Magpantheonsel-lark-v4x1.6.2RP-Cydonia-vXXX-22B-6  # Main writing engine
    parameters:
      weight: 0.72  # Increased to compensate for other reductions
      density: 0.78 # Increased for better detail

  # World Building & Character Focus (Enhanced)
  - model: Kaoeiri/MS-Inky-2409-22B  # World dynamics with character focus
    parameters:
      weight: 0.42  # Increased for stronger character-world integration
      density: 0.85 # Increased for richer character detail

  - model: Gryphe/Pantheon-RP-Pure-1.6.2-22b-Small  # Character interaction specialist
    parameters:
      weight: 0.48  # Increased for deeper character development
      density: 0.82 # Enhanced for personality detail

  # Combat & Magic (Rebalanced)
  - model: DigitalSouls/BlackSheep-DigitalSoul-22B  # Combat
    parameters:
      weight: 0.16  # Significantly reduced
      density: 0.70 # Maintained for quality

  - model: InferenceIllusionist/SorcererLM-22B  # Magical elements
    parameters:
      weight: 0.35  # Increased to maintain magical detail
      density: 0.78 # Enhanced for better integration

  # Story Elements (Adjusted)
  - model: TheDrummer/Cydonia-22B-v1.1
    parameters:
      weight: 0.22  # Slight increase for narrative depth
      density: 0.72

  - model: crestf411/MS-sunfall-v0.7.0  # Writing enhancement
    parameters:
      weight: 0.22  # Increased for better writing quality
      density: 0.75

  - model: Kaoeiri/MS_a-coolyte-2409-22B  # Character style content
    parameters:
      weight: 0.25  # Increased for better character elements
      density: 0.76

  # Ideas Generation
  - model: Kaoeiri/MS_fujin-2409-22B
    parameters:
      weight: 0.07  # Significantly reduced
      density: 0.62

  - model: Kaoeiri/MS_dampf-2409-22B
    parameters:
      weight: 0.12
      density: 0.68

  # Story Enhancement & Character Detail
  - model: hf-100/Mistral-Small-Spellbound-StoryWriter-22B-instruct-0.2-chkpt-200-16-bit
    parameters:
      weight: 0.28  # Increased for better story-character integration
      density: 0.75

  - model: ArliAI/Mistral-Small-22B-ArliAI-RPMax-v1.1  # Character interaction
    parameters:
      weight: 0.18  # Increased for better character dynamics
      density: 0.68

  - model: Darkknight535/MS-Moonlight-22B-v3  # Character detail specialist
    parameters:
      weight: 0.22  # Increased for deeper character profiles
      density: 0.72

  - model: Envoid/Mistral-Small-NovusKyver  # Maintained core functionality
    parameters:
      weight: 0.12  # Reduced to minimize instruction avoidance
      density: 0.65 # Adjusted down to reduce negative behaviors

  - model: TroyDoesAI/BlackSheep-MermaidMistral-22B  # Character moments
    parameters:
      weight: 0.25  # Slight increase for character interaction
      density: 0.74 # Increased for better detail

merge_method: dare_ties
base_model: unsloth/Mistral-Small-Instruct-2409
parameters:
  density: 0.98    # Increased for maximum character detail
  epsilon: 0.047   # Reduced for better consistency
  lambda: 1.58     # Increased for creative elements while maintaining coherence
dtype: bfloat16
tokenizer_source: union