Instructions to use NovaCorp/GRPO-RPG.System-3.2-1B-Chaos with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use NovaCorp/GRPO-RPG.System-3.2-1B-Chaos with Transformers:

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

pipe = pipeline("text-generation", model="NovaCorp/GRPO-RPG.System-3.2-1B-Chaos")

# Load model directly
from transformers import AutoTokenizer, AutoModelForMultimodalLM

tokenizer = AutoTokenizer.from_pretrained("NovaCorp/GRPO-RPG.System-3.2-1B-Chaos")
model = AutoModelForMultimodalLM.from_pretrained("NovaCorp/GRPO-RPG.System-3.2-1B-Chaos")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use NovaCorp/GRPO-RPG.System-3.2-1B-Chaos with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "NovaCorp/GRPO-RPG.System-3.2-1B-Chaos"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "NovaCorp/GRPO-RPG.System-3.2-1B-Chaos",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/NovaCorp/GRPO-RPG.System-3.2-1B-Chaos

SGLang

How to use NovaCorp/GRPO-RPG.System-3.2-1B-Chaos 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 "NovaCorp/GRPO-RPG.System-3.2-1B-Chaos" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "NovaCorp/GRPO-RPG.System-3.2-1B-Chaos",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "NovaCorp/GRPO-RPG.System-3.2-1B-Chaos" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "NovaCorp/GRPO-RPG.System-3.2-1B-Chaos",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use NovaCorp/GRPO-RPG.System-3.2-1B-Chaos with Docker Model Runner:
```
docker model run hf.co/NovaCorp/GRPO-RPG.System-3.2-1B-Chaos
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

GRPO RPG System 3 2 1B "Chaos"

Overview

GRPO RPG System 3.2 1B Chaos is the most aggressive and experimental member of the GRPO RPG System family.

Built from the fusion of Ultimate-RPG.System-3.2-1B and a GRPO-optimized conversational model, Chaos intentionally grants greater influence to the GRPO side of the merge while preserving the narrative foundations of RPG System.

The result is a model that prioritizes creativity, adaptability and unpredictability over strict conservatism. Chaos is designed to push the limits of a 1B architecture and explore more daring conversational and storytelling behaviors.

Architecture

Base architecture: Llama 3.2 1B
Parameters: 1B
Merge method: SLERP
Precision: FP16
RPG System influence: ~55%
GRPO influence: ~45%

Intended Behavior

GRPO RPG System 3.2 1B Chaos is designed for:

Experimental roleplay.
Unconventional storytelling.
Character improvisation.
Dynamic conversations.
Fantasy, science fiction and horror.
Alternate history and speculative worlds.
Creative writing with minimal constraints.

Chaos encourages:

High creativity.
Unusual ideas and scenarios.
Flexible tone adaptation.
Strong character expression.
Narrative spontaneity.

Strengths

Highly creative outputs.
Strong improvisational abilities.
Rich and expressive writing style.
Flexible dialogue generation.
Capable of surprising and unconventional responses.
Maintains good narrative capabilities despite its aggressive merge ratio.

Limitations

Experimental merge.
Less predictable than the Lite and Balanced variants.
Can occasionally sacrifice consistency in favor of creativity.
Long-context coherence remains constrained by the underlying 1B architecture.
Output quality may vary significantly depending on prompt style and inference settings.

Recommended Settings

Temperature: 1.1 – 1.5
Top-p: 0.92 – 0.99
Min-p: 0.03 – 0.08
Repetition penalty: 1.03 – 1.12
Context: 8K recommended

Philosophy

Chaos was created with a simple objective:

Explore the limits of creativity and conversational freedom within a compact 1B architecture.

It is not the safest member of the family. It is not the most predictable.

Chaos embraces experimentation.

Sometimes brilliant. Sometimes strange.

Always interesting.

Family

Lite → Conservative and stable.
Balanced → Equilibrium between narrative depth and conversational flexibility.
Chaos → Aggressive, creative and experimental.

Version

GRPO RPG System 3.2 1B Chaos

Aggressive Variant — Maximum creativity, stronger GRPO influence and experimental behavior.

For those who prefer exploration over certainty.

Models Merged

The following models were included in the merge:

Configuration

The following YAML configuration was used to produce this model:



# Author: Dr. Novaciano
# Objective: GRPO RPG Unethic 3.2 1B AI Model
# =========================================================
# PROJECT: GRPO RPG System 3.2 1B
# =========================================================

models:
  - model: NovaCorp/Ultimate-RPG.System-3.2-1B  # Experimental viral strain neural imprint
  - model: jtatman/llama3.2_1b_uncensored_pentest_grpo-merged  # Baseline cognitive template, "safe mode"

merge_method: slerp  # Spherical Linear Interpolation to preserve extreme viral traits smoothly
base_model: NovaCorp/Ultimate-RPG.System-3.2-1B  # Anchor model for stable latent space

dtype: bfloat16  # Memory-efficient precision, minimal loss in viral feature fidelity

parameters:
  t: 0.45
  normalize: false
  rescale: true
  rescale_factor: 1.12
  memory_efficient: true
  low_cpu_mem_usage: true

layer_range:
  - value: [4, 22]

tie_word_embeddings: false
tie_output_embeddings: false

Downloads last month: 43

Safetensors

Model size

1B params

Tensor type

BF16

Model tree for NovaCorp/GRPO-RPG.System-3.2-1B-Chaos

NovaCorp/Ultimate-RPG.System-3.2-1B

jtatman/llama3.2_1b_uncensored_pentest_grpo-merged

Merge model

this model

Quantizations

2 models