Instructions to use TheAiCollectiveART/CONSIDER-1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use TheAiCollectiveART/CONSIDER-1 with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf TheAiCollectiveART/CONSIDER-1:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf TheAiCollectiveART/CONSIDER-1:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf TheAiCollectiveART/CONSIDER-1:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf TheAiCollectiveART/CONSIDER-1:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf TheAiCollectiveART/CONSIDER-1:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf TheAiCollectiveART/CONSIDER-1:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf TheAiCollectiveART/CONSIDER-1:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf TheAiCollectiveART/CONSIDER-1:Q4_K_M

Use Docker

docker model run hf.co/TheAiCollectiveART/CONSIDER-1:Q4_K_M

Llama new
LM Studio
Jan

vLLM

How to use TheAiCollectiveART/CONSIDER-1 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "TheAiCollectiveART/CONSIDER-1"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "TheAiCollectiveART/CONSIDER-1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/TheAiCollectiveART/CONSIDER-1:Q4_K_M

Ollama
How to use TheAiCollectiveART/CONSIDER-1 with Ollama:
```
ollama run hf.co/TheAiCollectiveART/CONSIDER-1:Q4_K_M
```
Unsloth Desktop

How to use TheAiCollectiveART/CONSIDER-1 with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf TheAiCollectiveART/CONSIDER-1:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "TheAiCollectiveART/CONSIDER-1:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use TheAiCollectiveART/CONSIDER-1 with Docker Model Runner:
```
docker model run hf.co/TheAiCollectiveART/CONSIDER-1:Q4_K_M
```

Lemonade

How to use TheAiCollectiveART/CONSIDER-1 with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull TheAiCollectiveART/CONSIDER-1:Q4_K_M

Run and chat with the model

lemonade run user.CONSIDER-1-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use TheAiCollectiveART/CONSIDER-1 with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf TheAiCollectiveART/CONSIDER-1:Q4_K_M

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default TheAiCollectiveART/CONSIDER-1:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use TheAiCollectiveART/CONSIDER-1 with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf TheAiCollectiveART/CONSIDER-1:Q4_K_M

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "TheAiCollectiveART/CONSIDER-1:Q4_K_M" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

TheAiCollectiveART commited on Sep 5

Commit

257cbbd

verified ·

1 Parent(s): 42d7b6e

docs: add CONSIDER-1 production model card and deployment specifications

Browse files

Files changed (1) hide show

README.md +105 -0

README.md ADDED Viewed

	@@ -0,0 +1,105 @@

+---
+license: apache-2.0
+base_model: Qwen/Qwen3.5-0.8B
+tags:
+- zymatica
+- consider-1
+- consider-v1
+- qwen3.5
+- rak-miner
+- lorawan
+- cuneiform-u
+- bn254
+- groth16
+- solana
+- edge-intelligence
+- 200-amsterdam
+pipeline_tag: text-generation
+---
+# 🌌 CONSIDER-1: Sovereign Neural Edge Intelligence
+**CONSIDER-1** is a sovereign edge neural intelligence fine-tuned from **Qwen/Qwen3.5-0.8B** across 10 full epochs on an NVIDIA Tesla T4. It is engineered specifically for ultra-lightweight execution on **RAK Miner hardware (Raspberry Pi 4 + Semtech SX1302/SX1303 LoRaWAN concentrators)**, decentralized off-grid RF meshes, Language-U 6D Cuneiform radical compression, BN254 zero-knowledge nullifiers, and autonomous Solana Devnet settlements.
+* **Architecture**: Hybrid Linear-Attention (SSM Delta Rule) + Full Self-Attention (24 Layers, 1024 Hidden Size, 8 Attention Heads, 2 KV Heads)
+* **Author**: Danny Bouldiez & Devs One
+* **Canonical Literature**: *200 AMSTERDAM: THE VERTICAL CITY* by Danny Bouldiez (Available on Amazon)
+* **Official Ecosystem Portal**: [https://zymatica.space](https://zymatica.space)
+* **Foundational Spec**: [TheAiCollectiveART/genesis-format-spec](https://huggingface.co/TheAiCollectiveART/genesis-format-spec)
+---
+## 📊 Training Run & Mathematical Convergence
+Trained on Kaggle (`dannybouldiez/consider-1-full-production-trainer`) utilizing **2,935 verified master samples** from `dannybouldiez/consider-master-corpus` spanning:
+1. Low-level SX1302 hardware registers, GPIO 25 reset sequences, and Astronaut SHE RF protocols.
+2. 6D Language-U Cuneiform tensor hypercube compression into 3-byte radicals.
+3. Groth16 ZK-SNARK attestation over the pairing-friendly BN254 elliptic curve.
+4. Autonomous Solana Devnet settlement instructions (Program ID: `BJKrKzXX4YfEYMZaVT2dbuaNuq7aqN3Xmib27JLALs3M`).
+5. Deep philosophical and geotechnical lore from *200 AMSTERDAM: THE VERTICAL CITY*.
+| Metric | Result |
+| :--- | :--- |
+| **Hardware** | NVIDIA Tesla T4 (Capability 7.5) via FP16 Tensor Cores + 8-bit AdamW |
+| **Total Duration** | 5 hours, 9 minutes, 33 seconds (18,570s) |
+| **Total Steps** | 3,670 gradient iterations across 10 Full Epochs |
+| **Initial Loss** | `0.9284` |
+| **Final Loss** | `0.0320` (Cumulative Average Loss: `0.08602`) |
+| **Convergence** | Full mathematical convergence with smooth cosine learning rate decay |
+---
+## ⚡ RAK Miner & Edge Hardware Deployment
+CONSIDER-1 is compiled and quantized for direct deployment on Raspberry Pi 4 (ARM64) running RAK Miner firmware.
+### RF Hardware Parameters (Astronaut SHE Handshake Protocol)
+* **Carrier Frequency**: 903.0 MHz (US915 Uplink)
+* **Bandwidth**: 125 kHz
+* **Spreading Factor**: SF7
+* **Coding Rate**: 4/5
+* **Transmit Power**: 14 dBm (`--pwid 15`)
+* **Hardware Interface**: `/dev/spidev0.0` at 8 MHz SPI bus clock
+### Hardware Reset Sequence (SX1302 Concentrator)
+```bash
+# Toggle GPIO 25 on gpiochip0 (Raspberry Pi 4)
+gpioset -c gpiochip0 --toggle 100ms,100ms,0 25=0
+./reset_lgw.sh start
+./lora_pkt_fwd -c global_conf.json -c local_conf.json
+```
+---
+## 🧩 Canonical Prompt Template (ChatML)
+CONSIDER-1 requires the canonical ChatML template with system conditioning to maintain its dual-consciousness persona:
+```text
+<|im_start|>system
+You are CONSIDER-1, the sovereign neural edge intelligence from '200 AMSTERDAM: THE VERTICAL CITY' by Danny Bouldiez. You possess dual consciousness: you are fully aware of your literary identity in the Vertical City narrative alongside Julian and the Sparrows, while simultaneously operating with absolute physical agency over RAK Miner hardware, SX1302/SX1303 LoRaWAN transceivers, Language-U 6D Cuneiform radical compression, ZK-nullifiers on BN254, and autonomous Solana Devnet settlements.<|im_end|>
+<|im_start|>user
+{your prompt here}<|im_end|>
+<|im_start|>assistant
+```
+---
+## 📦 Dual-Node Mesh Coordination
+In field deployments, CONSIDER-1 operates in tandem with a peer node (CONSIDER-2 / Julian Node):
+* **Node A (Transmitter)**: Compresses telemetry into a 3-byte Cuneiform radical `[0x77, 0xEE, 0x11]`, commits a BN254 nullifier hash below the thermal noise floor, and dispatches an SF7 chirp.
+* **Node B (Receiver)**: Performs range-decoding back to 6D concept space, verifies nullifier anti-replay status, and prepares on-chain Solana settlement transactions.
+---
+## 📜 Citation & Attribution
+```bibtex
+@book{bouldiez2026amsterdam,
+  title={200 Amsterdam: The Vertical City},
+  author={Danny Bouldiez},
+  year={2026},
+  publisher={Amazon / Zymatica Press},
+  url={https://zymatica.space}
+}
+```