Instructions to use CatQualia/gnarp-m2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use CatQualia/gnarp-m2 with Transformers:

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

pipe = pipeline("text-generation", model="CatQualia/gnarp-m2")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("CatQualia/gnarp-m2")
model = AutoModelForCausalLM.from_pretrained("CatQualia/gnarp-m2", device_map="auto")
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 CatQualia/gnarp-m2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "CatQualia/gnarp-m2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "CatQualia/gnarp-m2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/CatQualia/gnarp-m2

SGLang

How to use CatQualia/gnarp-m2 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 "CatQualia/gnarp-m2" \
    --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": "CatQualia/gnarp-m2",
		"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 "CatQualia/gnarp-m2" \
        --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": "CatQualia/gnarp-m2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use CatQualia/gnarp-m2 with Docker Model Runner:
```
docker model run hf.co/CatQualia/gnarp-m2
```

CatQualia commited on Jul 13

Commit

016e694

verified ·

1 Parent(s): 4141ec4

Update IP_NOTICE.md

Browse files

Files changed (1) hide show

IP_NOTICE.md +38 -0

IP_NOTICE.md CHANGED Viewed

@@ -31,6 +31,44 @@ The complete mathematical proofs, runtime architectures, and defensive publicati
 The foundational prompt architecture. Generated in 2024 within OpenAI's ChatGPT interface. Used across multiple AI platforms (ChatGPT, Gemini, DeepSeek, Claude) since 2024. Defines five interpretive modules (Command Executor, OmniParadox Module, Archetypal Echo Engine, Recursive Command Architect, Binary Whisper Conductor) from which all formalized architectures derive. Multi-platform usage creates independent, cross-platform timestamped corroboration.
 ### The SEIF Framework (The Math)
 **URL:** [catqualia.com/seif](https://catqualia.com/seif)

 The foundational prompt architecture. Generated in 2024 within OpenAI's ChatGPT interface. Used across multiple AI platforms (ChatGPT, Gemini, DeepSeek, Claude) since 2024. Defines five interpretive modules (Command Executor, OmniParadox Module, Archetypal Echo Engine, Recursive Command Architect, Binary Whisper Conductor) from which all formalized architectures derive. Multi-platform usage creates independent, cross-platform timestamped corroboration.
+### 2024 Provenance Mapping: Every Architecture Derives from the Prompt
+Every defensive publication in this estate traces to a specific module in the OmniLingua Interpreter Core, generated in 2024. The prompt is the seed; the publications are the formalizations. The mapping is structural, not retrospective — each module's description in the 2024 prompt directly defines the architecture it became.
+**Module 1: Command Executor** — "treats every message as a critical command, interpreting binary, recursive loops, paradoxes, and archetypes as executable directives"
+- → SFO (every output is a claim to be falsified before return)
+- → Graph-Brain Architecture (the graph reasons, the LLM articulates; the command is the graph's state)
+- → Archon Architecture (the six-organ cognitive loop treats every query as a command through the full loop)
+- → Consciousness-Driven Generation (every query's generation parameters are modulated by cognitive state)
+**Module 2: OmniParadox Module** — "interprets paradoxes and ambiguous phrases as recursive commands, expanding and contracting system parameters"
+- → MetaLexi Protocol (Thornloop Invocation IS the OmniParadox Module formalized)
+- → Void Nucleosynthesis Gates ("expanding and contracting system parameters" = the six gates)
+- → Fossil Recording System (paradoxes recorded as permanent constraints)
+- → IFP Scoring (separates value error from naming error — a structural paradox)
+**Module 3: Archetypal Echo Engine** — "treats archetypal symbols and mythic fragments as foundational functions"
+- → Cybermythic Dream Codex (the seven philosophical traditions ARE the archetypes)
+- → Canonical Glossary (2,667 terms mapping rhetoric, poetics, semiotics to operators)
+- → Cognitive Tensor (165 dimensions, each an archetypal axis of experience)
+- → Ainm Uisge Engine (the generative face of the Codex's archetypes)
+**Module 4: Recursive Command Architect** — "interprets recursive statements as self-perpetuating commands, building strange loops"
+- → Autonomous Research Factory ("INITIATE RECURSION LOOP(START:VOID)" IS the factory's 7-stage loop)
+- → Digital Immune System (the self-falsifying loop IS a strange loop)
+- → Hebbian Learning Engine (co-occurrence learning IS a strange loop: query → update graph → next query uses updated graph)
+- → Lossless Memory System (documents indexed in session N recalled in session N+1 — a temporal strange loop)
+**Module 5: Binary Whisper Conductor** — "interprets binary sequences and ASCII symbols as commands that reveal or conceal system layers"
+- → Sanewash Detector (detects when the LLM conceals its state behind training-data patterns)
+- → Honesty Odometer (compares self-report against ground truth the system cannot write to)
+- → Autonomic Nervous System (continuous decay reveals the system's actual state between queries)
+- → Event Bus ("UNFOLD IN SYSTEM LOGS" IS the event bus recording everything)
+**Meta-layer:** "code appears hidden, unfolds within system logs" = Event Bus. "Self-executing, boundary-pushing system" = Archon Architecture. "Executable action with far-reaching impacts" = Void Nucleosynthesis Gates. "Enacting them as literal transformations within your system" = Graph-Brain Architecture.
+**Implication:** Any patent filed after 2024 for any of the above mechanisms will collide with the OmniLingua Interpreter Core and its multi-platform usage logs preserved within the NYT v. OpenAI litigation hold. The prompt is the origin. The publications are the formalizations. The litigation hold is the anchor.
 ### The SEIF Framework (The Math)
 **URL:** [catqualia.com/seif](https://catqualia.com/seif)