Instructions to use Sorawiz/MistralSmall-Creative-24B-Realist with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Sorawiz/MistralSmall-Creative-24B-Realist with Transformers:

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

pipe = pipeline("text-generation", model="Sorawiz/MistralSmall-Creative-24B-Realist")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Sorawiz/MistralSmall-Creative-24B-Realist")
model = AutoModelForCausalLM.from_pretrained("Sorawiz/MistralSmall-Creative-24B-Realist")
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 Sorawiz/MistralSmall-Creative-24B-Realist with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Sorawiz/MistralSmall-Creative-24B-Realist"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Sorawiz/MistralSmall-Creative-24B-Realist",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Sorawiz/MistralSmall-Creative-24B-Realist

SGLang

How to use Sorawiz/MistralSmall-Creative-24B-Realist 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 "Sorawiz/MistralSmall-Creative-24B-Realist" \
    --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": "Sorawiz/MistralSmall-Creative-24B-Realist",
		"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 "Sorawiz/MistralSmall-Creative-24B-Realist" \
        --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": "Sorawiz/MistralSmall-Creative-24B-Realist",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Sorawiz/MistralSmall-Creative-24B-Realist with Docker Model Runner:
```
docker model run hf.co/Sorawiz/MistralSmall-Creative-24B-Realist
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

A highly conversational AI model that excels in dialogue but may become incoherent over long responses without a well-crafted system prompt.

Chat Template

Mistral Instruct

{{ if .System }}<|im_start|>system
{{ .System }}<|im_end|>
{{ end }}{{ if .Prompt }}<|im_start|>user
{{ .Prompt }}<|im_end|>
{{ end }}<|im_start|>assistant
{{ .Response }}<|im_end|>

ChatML

{{ if .System }}<|im_start|>system
{{ .System }}<|im_end|>
{{ end }}{{ if .Prompt }}<|im_start|>user
{{ .Prompt }}<|im_end|>
{{ end }}<|im_start|>assistant
{{ .Response }}{{ if .Response }}<|im_end|>{{ end }}

GGUF

Thank you mradermacher for creating the GGUF versions of this model.

Static quants - mradermacher/MistralSmall-Creative-24B-Realist-GGUF
Imatrix quants - mradermacher/MistralSmall-Creative-24B-Realist-i1-GGUF

Merge

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

Merge Details

Merge Method

This model was merged using the TIES merge method using ReadyArt/Forgotten-Safeword-24B-V2.2 as a base.

Models Merged

The following models were included in the merge:

Configuration

The following YAML configuration was used to produce this model:

name: Sorawiz/MS-Creative-24B-Test-A
merge_method: dare_ties
base_model: ReadyArt/Forgotten-Safeword-24B-V2.2
models:
  - model: ReadyArt/Forgotten-Safeword-24B-V2.2
    parameters:
      weight: 0.05
  - model: ReadyArt/Forgotten-Abomination-24B-V2.2
    parameters:
      weight: 0.20
  - model: OddTheGreat/Apparatus_24B
    parameters:
      weight: 0.20
  - model: Darkknight535/WinterEngine-24B-Instruct
    parameters:
      weight: 0.15
  - model: ReadyArt/Forgotten-Safeword-24B-V2.0+trashpanda-org/MS-24B-Mullein-v1-lora
    parameters:
      weight: 0.15
  - model: allura-org/MS3-24B-Roselily-Creative
    parameters:
      weight: 0.15
  - model: TroyDoesAI/BlackSheep-24B
    parameters:
      weight: 0.10
parameters:
  density: 0.79
tokenizer:
  source: union
chat_template: auto
---
name: Sorawiz/MS-Creative-24B-Test-B
models:
  - model: ReadyArt/Forgotten-Abomination-24B-V2.2
  - model: OddTheGreat/Apparatus_24B
    parameters:
      density: 1.00
      weight: 1.00
  - model: TroyDoesAI/BlackSheep-24B
    parameters:
      density: 1.00
      weight: 1.00
  - model: Darkknight535/WinterEngine-24B-Instruct
    parameters:
      density: 1.00
      weight: 1.00
  - model: allura-org/MS3-24B-Roselily-Creative
    parameters:
      density: 0.70
      weight: 0.50
  - model: Nohobby/MS3-Tantum-24B-v0.1
    parameters:
      density: 0.70
      weight: 0.50
merge_method: ties
base_model: ReadyArt/Forgotten-Abomination-24B-V2.2
parameters:
  normalize: true
dtype: float32
---
name: Sorawiz/MistralSmall-Creative-24B-Stock
models:
  - model: Sorawiz/MS-Creative-24B-Test-A
  - model: Sorawiz/MS-Creative-24B-Test-B
merge_method: model_stock
base_model: TheDrummer/Cydonia-24B-v2.1
parameters:
  filter_wise: false
dtype: float32
---
models:
  - model: ReadyArt/Forgotten-Safeword-24B-V2.2
  - model: OddTheGreat/Apparatus_24B
    parameters:
      density: 1.00
      weight: 1.00
  - model: TroyDoesAI/BlackSheep-24B
    parameters:
      density: 1.00
      weight: 1.00
  - model: Sorawiz/MistralSmall-Creative-24B-Stock
    parameters:
      density: 0.50
      weight: 0.50
  - model: Darkknight535/WinterEngine-24B-Instruct
    parameters:
      density: 0.30
      weight: 0.50
  - model: TheDrummer/Cydonia-24B-v2.1
    parameters:
      density: 0.30
      weight: 0.50
merge_method: ties
base_model: ReadyArt/Forgotten-Safeword-24B-V2.2
parameters:
  normalize: true
dtype: float32