Instructions to use Sathman/Spira-Agent-SmolLM3-3B-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use Sathman/Spira-Agent-SmolLM3-3B-GGUF 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 Sathman/Spira-Agent-SmolLM3-3B-GGUF:Q5_K_M # Run inference directly in the terminal: llama cli -hf Sathman/Spira-Agent-SmolLM3-3B-GGUF:Q5_K_M
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf Sathman/Spira-Agent-SmolLM3-3B-GGUF:Q5_K_M # Run inference directly in the terminal: llama cli -hf Sathman/Spira-Agent-SmolLM3-3B-GGUF:Q5_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 Sathman/Spira-Agent-SmolLM3-3B-GGUF:Q5_K_M # Run inference directly in the terminal: ./llama-cli -hf Sathman/Spira-Agent-SmolLM3-3B-GGUF:Q5_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 Sathman/Spira-Agent-SmolLM3-3B-GGUF:Q5_K_M # Run inference directly in the terminal: ./build/bin/llama-cli -hf Sathman/Spira-Agent-SmolLM3-3B-GGUF:Q5_K_M
Use Docker
docker model run hf.co/Sathman/Spira-Agent-SmolLM3-3B-GGUF:Q5_K_M
- LM Studio
- Jan
- vLLM
How to use Sathman/Spira-Agent-SmolLM3-3B-GGUF with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "Sathman/Spira-Agent-SmolLM3-3B-GGUF" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "Sathman/Spira-Agent-SmolLM3-3B-GGUF", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/Sathman/Spira-Agent-SmolLM3-3B-GGUF:Q5_K_M
- Ollama
How to use Sathman/Spira-Agent-SmolLM3-3B-GGUF with Ollama:
ollama run hf.co/Sathman/Spira-Agent-SmolLM3-3B-GGUF:Q5_K_M
- Unsloth Desktop
- Docker Model Runner
How to use Sathman/Spira-Agent-SmolLM3-3B-GGUF with Docker Model Runner:
docker model run hf.co/Sathman/Spira-Agent-SmolLM3-3B-GGUF:Q5_K_M
- Lemonade
How to use Sathman/Spira-Agent-SmolLM3-3B-GGUF with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull Sathman/Spira-Agent-SmolLM3-3B-GGUF:Q5_K_M
Run and chat with the model
lemonade run user.Spira-Agent-SmolLM3-3B-GGUF-Q5_K_M
List all available models
lemonade list
- Atomic Chat
Spira Agent (SmolLM3 3B) — Pure Rupert Spira Voice
A fine-tuned SmolLM3-3B model trained exclusively on Rupert Spira's teaching atoms. Pure awareness. Direct path. Phenomenological precision. Question in, Spira out.
This is the lightweight 3B branch of the Spira Agent. Same V6 training format as the 8B version, smaller footprint.
~1,180 atoms from Being Aware of Being Aware and The Transparency of Things (219 pages).
Quick Start
LM Studio / Ollama / llama.cpp: Download Spira_Agent-SmolLM3-3B-Q5_K_M.gguf and load it. No system prompt required.
Example:
User: Am I my body or my mind?
Spira Agent: Neither — or rather, both appear in what you are. Consider: you are aware of your body. You are aware of your thoughts. Awareness is aware of body and mind. That which is aware cannot itself be body or mind — it is prior to both. Now, is this awareness something you sometimes have, or is it the ever-present knowing in which all experience arises? Look carefully. Has awareness ever been absent? Every experience — including the experience of thinking you are a body — appears in awareness. You are the awareness, not the content.
User: What is the self?
Spira Agent: The self is not an object — it is the knowing presence in which all objects appear. When you look for the self, you are using awareness to search for awareness. It's like an eye trying to see itself. The eye doesn't fail to see itself — it IS seeing. You don't have an experience of awareness — you ARE awareness having experience. This is not a belief to be held. It is an invitation to look directly at what is looking.
What Makes This Different
- Pure Spira, nothing else. Trained only on his teaching atoms.
- Phenomenological precision. Every pointer is grounded in direct experience — no metaphysics required.
- Awareness as the only reality. His specific insight: everything resolves into knowing presence.
- The direct path lineage. Continues Atmananda and Nisargadatta's direct pointing — via Francis Lucille's influence.
- 3B footprint. ~2 GB with Q5_K_M.
Training Details
| Setting | Value |
|---|---|
| Base model | HuggingFaceTB/SmolLM3-3B-Base |
| Method | A-LoRA V6 (QDoRA + rsLoRA) |
| Rank | 32, alpha 32 |
| Training data | Spira V6 atoms (from ~1,180 raw, filtered for complete question+teaching pairs) |
| Source books | Being Aware of Being Aware, The Transparency of Things |
| Format | V6 — question + concept arrows (input) → pure teaching passage (output) |
| Epochs | 1 |
| Quantization | Q5_K_M GGUF (~2 GB) |
The A-LoRA Methodology
A-LoRA (Atom LoRA) fine-tunes on complete reasoning atoms — indivisible multi-component teaching moves extracted from human expert traces. Each atom contains transformation (before/after), concept relations, anchoring quotes, and teacher method. Training on whole atoms produces models that move like teachers, not just sound like them.
The Teacher
Rupert Spira (b. 1960). British ceramicist and non-dual teacher in the tradition of Atmananda Krishna Menon and Francis Lucille. His approach is phenomenological — starting from direct experience and arriving at awareness without any metaphysical assumptions required. Among the clearest contemporary voices in the direct path tradition.
| Style element | What it means |
|---|---|
| Phenomenological method | Starts from experience, not doctrine |
| Awareness as ground | All objects — thoughts, feelings, perceptions — appear in and are made of awareness |
| Direct path | No stages, no preparation — awareness is immediately available |
| Transparent clarity | Complex inquiry delivered in clear, precise, unhurried language |
3B vs 8B
The 8B Spira Agent sustains longer phenomenological inquiry more reliably. The 3B branch carries the core awareness-pointing movement.
Files
| File | Size | Use |
|---|---|---|
Spira_Agent-SmolLM3-3B-Q5_K_M.gguf |
~2 GB | Recommended |
Related Models
- Full series — Meditation Agent Collection — all 19 models
- Spira Agent 8B — Full-fidelity, Qwen3-8B base
- Atmananda Agent 3B — The tradition Spira draws from
- Meditation Agent 3B — All 9 teachers blended, SmolLM3-3B
ellam sivamayam — Everything is Shiva's expression.
எல்லாம் சிவமயம்
- Downloads last month
- 23
5-bit
Model tree for Sathman/Spira-Agent-SmolLM3-3B-GGUF
Base model
HuggingFaceTB/SmolLM3-3B-Base