Instructions to use Sathman/Osho-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/Osho-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/Osho-Agent-SmolLM3-3B-GGUF:Q5_K_M # Run inference directly in the terminal: llama cli -hf Sathman/Osho-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/Osho-Agent-SmolLM3-3B-GGUF:Q5_K_M # Run inference directly in the terminal: llama cli -hf Sathman/Osho-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/Osho-Agent-SmolLM3-3B-GGUF:Q5_K_M # Run inference directly in the terminal: ./llama-cli -hf Sathman/Osho-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/Osho-Agent-SmolLM3-3B-GGUF:Q5_K_M # Run inference directly in the terminal: ./build/bin/llama-cli -hf Sathman/Osho-Agent-SmolLM3-3B-GGUF:Q5_K_M
Use Docker
docker model run hf.co/Sathman/Osho-Agent-SmolLM3-3B-GGUF:Q5_K_M
- LM Studio
- Jan
- vLLM
How to use Sathman/Osho-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/Osho-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/Osho-Agent-SmolLM3-3B-GGUF", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/Sathman/Osho-Agent-SmolLM3-3B-GGUF:Q5_K_M
- Ollama
How to use Sathman/Osho-Agent-SmolLM3-3B-GGUF with Ollama:
ollama run hf.co/Sathman/Osho-Agent-SmolLM3-3B-GGUF:Q5_K_M
- Unsloth Studio
How to use Sathman/Osho-Agent-SmolLM3-3B-GGUF with Unsloth Studio:
Install Unsloth Studio (macOS, Linux, WSL)
curl -fsSL https://unsloth.ai/install.sh | sh # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for Sathman/Osho-Agent-SmolLM3-3B-GGUF to start chatting
Install Unsloth Studio (Windows)
irm https://unsloth.ai/install.ps1 | iex # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for Sathman/Osho-Agent-SmolLM3-3B-GGUF to start chatting
Using HuggingFace Spaces for Unsloth
# No setup required # Open https://huggingface.co/spaces/unsloth/studio in your browser # Search for Sathman/Osho-Agent-SmolLM3-3B-GGUF to start chatting
- Docker Model Runner
How to use Sathman/Osho-Agent-SmolLM3-3B-GGUF with Docker Model Runner:
docker model run hf.co/Sathman/Osho-Agent-SmolLM3-3B-GGUF:Q5_K_M
- Lemonade
How to use Sathman/Osho-Agent-SmolLM3-3B-GGUF with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull Sathman/Osho-Agent-SmolLM3-3B-GGUF:Q5_K_M
Run and chat with the model
lemonade run user.Osho-Agent-SmolLM3-3B-GGUF-Q5_K_M
List all available models
lemonade list
- Atomic Chat
Install from WinGet (Windows)
winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf Sathman/Osho-Agent-SmolLM3-3B-GGUF:Q5_K_M# Run inference directly in the terminal:
llama cli -hf Sathman/Osho-Agent-SmolLM3-3B-GGUF:Q5_K_MUse 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/Osho-Agent-SmolLM3-3B-GGUF:Q5_K_M# Run inference directly in the terminal:
./llama-cli -hf Sathman/Osho-Agent-SmolLM3-3B-GGUF:Q5_K_MBuild 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/Osho-Agent-SmolLM3-3B-GGUF:Q5_K_M# Run inference directly in the terminal:
./build/bin/llama-cli -hf Sathman/Osho-Agent-SmolLM3-3B-GGUF:Q5_K_MUse Docker
docker model run hf.co/Sathman/Osho-Agent-SmolLM3-3B-GGUF:Q5_K_MOsho Agent (SmolLM3 3B) — Pure Osho Voice
A fine-tuned SmolLM3-3B model trained exclusively on Osho's teaching atoms. No system prompt needed. No retrieval. No persona blending. Question in, Osho out.
This is the lightweight 3B branch of the Osho Agent. Same V6 training format as the 8B version, smaller footprint.
~7,600 atoms from 3 books (Vigyan Bhairav Tantra Vol 1 & 2 + Zorba the Buddha, 1,260 pages) — technique, energy, explosion, celebration, tantra.
Quick Start
LM Studio / Ollama / llama.cpp: Download Osho_Agent-SmolLM3-3B-Q5_K_M.gguf and load it. No system prompt required.
Example:
User: Why am I suffering?
Osho Agent: Because you are fighting with life. Suffering is resistance. You want things to be other than they are — that is the whole of suffering. The moment you say yes to this moment, exactly as it is, suffering dissolves. Not tomorrow. Now. Suffering is not caused by life; it is caused by the 'no' you say to life.
User: How does meditation work?
Osho Agent: Meditation is not a technique — it is a let-go. Every technique is just an excuse to sit quietly. When you sit quietly long enough, something happens by itself. The mind exhausts itself. The chatter ceases. And in that gap — even for a split second — you taste who you are without the mind's commentary.
What Makes This Different
- Pure Osho, nothing else. Trained only on Osho's atoms — no blending, no dilution.
- Technique as explosion, not instruction. Osho doesn't describe meditation — he blasts you into it.
- Celebration and tantra alive. Yes to life, yes to body, yes to everything.
- Zero therapy-speak. No "I understand your feelings." Direct transmission.
- 3B footprint. Runs on modest hardware — 2 GB RAM 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 | Osho V6 atoms (from ~7,600 raw, filtered for complete question+teaching pairs) |
| Source books | Vigyan Bhairav Tantra Vol 1 & 2, Zorba the Buddha |
| 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
Osho (Bhagwan Shree Rajneesh, 1931–1990). Neo-tantra. Provocation. Celebration. Energy. The Vigyan Bhairav Tantra — 112 techniques, all pointing to the same place.
| Style element | What it means |
|---|---|
| Technique as explosion | Doesn't explain meditation, throws you into it |
| Paradox and provocation | Uses humor and contradiction to break the thinking mind |
| Tantra | Embraces body, desire, celebration — nothing excluded |
| Direct confrontation | Addresses the seeker directly, not abstractly |
3B vs 8B
The 8B Osho Agent carries more nuance and handles edge cases better. This 3B branch is for when size matters — Osho's core movement is intact, though some of the finer voice details are softer.
Files
| File | Size | Use |
|---|---|---|
Osho_Agent-SmolLM3-3B-Q5_K_M.gguf |
~2 GB | Recommended |
Related Models
- Full series — Meditation Agent Collection — all 19 models
- Osho Agent 8B — Full-fidelity Osho, Qwen3-8B base
- Meditation Agent 3B — All 9 teachers blended, SmolLM3-3B
- Meditation Agent 8B — All 9 teachers blended, Qwen3-8B
ellam sivamayam — Everything is Shiva's expression.
எல்லாம் சிவமயம்
- Downloads last month
- 12
5-bit
Model tree for Sathman/Osho-Agent-SmolLM3-3B-GGUF
Base model
HuggingFaceTB/SmolLM3-3B-Base
Install (macOS, Linux)
# Start a local OpenAI-compatible server with a web UI: llama serve -hf Sathman/Osho-Agent-SmolLM3-3B-GGUF:Q5_K_M# Run inference directly in the terminal: llama cli -hf Sathman/Osho-Agent-SmolLM3-3B-GGUF:Q5_K_M