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

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use RegeneratusLabs/augury-1b 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 RegeneratusLabs/augury-1b:F16
# Run inference directly in the terminal:
llama cli -hf RegeneratusLabs/augury-1b:F16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf RegeneratusLabs/augury-1b:F16
# Run inference directly in the terminal:
llama cli -hf RegeneratusLabs/augury-1b:F16

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 RegeneratusLabs/augury-1b:F16
# Run inference directly in the terminal:
./llama-cli -hf RegeneratusLabs/augury-1b:F16

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 RegeneratusLabs/augury-1b:F16
# Run inference directly in the terminal:
./build/bin/llama-cli -hf RegeneratusLabs/augury-1b:F16

Use Docker

docker model run hf.co/RegeneratusLabs/augury-1b:F16

LM Studio
Jan

vLLM

How to use RegeneratusLabs/augury-1b with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/RegeneratusLabs/augury-1b:F16

Ollama
How to use RegeneratusLabs/augury-1b with Ollama:
```
ollama run hf.co/RegeneratusLabs/augury-1b:F16
```

Unsloth Studio

How to use RegeneratusLabs/augury-1b 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 RegeneratusLabs/augury-1b 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 RegeneratusLabs/augury-1b to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for RegeneratusLabs/augury-1b to start chatting

How to use RegeneratusLabs/augury-1b with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf RegeneratusLabs/augury-1b:F16

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "RegeneratusLabs/augury-1b:F16"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

OpenClaw new

How to use RegeneratusLabs/augury-1b with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf RegeneratusLabs/augury-1b:F16

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 "RegeneratusLabs/augury-1b:F16" \
  --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"

Docker Model Runner
How to use RegeneratusLabs/augury-1b with Docker Model Runner:
```
docker model run hf.co/RegeneratusLabs/augury-1b:F16
```

Lemonade

How to use RegeneratusLabs/augury-1b with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull RegeneratusLabs/augury-1b:F16

Run and chat with the model

lemonade run user.augury-1b-F16

List all available models

lemonade list

Hermes Agent

How to use RegeneratusLabs/augury-1b with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf RegeneratusLabs/augury-1b:F16

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 RegeneratusLabs/augury-1b:F16

Run Hermes

hermes

Atomic Chat

AUGURY 1B — weeds as soil indicators, in the farmer's voice

A LoRA fine-tune of MiniCPM5-1B that turns structured soil-indicator facts into a clear, conversational soil story. The model never invents facts — it receives them from the AUGURY species database and presents them.

species + indicators (from the AUGURY database)
        │  [MiniCPM5-1B + LoRA — this model]
        ▼
"This is what Capeweed is telling you about your soil. ..."

What it does (and doesn't)

✅ Presents given indicator facts in natural, farmer-friendly language
✅ Handles AU + EU species, region-aware phrasing
✅ Politely refuses non-plant / management questions
❌ Does not identify plants from photos (that's the retrieval layer)
❌ Does not generate indicator facts from memory
❌ Does not give management, herbicide, or remediation advice

Quick start

llama.cpp (GGUF — phone / edge)

wget https://huggingface.co/RegeneratusLabs/augury-1b/resolve/main/MiniCPM5-1B-AUGURY-Q4_K_M.gguf
llama-server -m MiniCPM5-1B-AUGURY-Q4_K_M.gguf --port 8080

The chat template (chatml, minicpm) is embedded in the GGUF. Feed it the system prompt + a user message shaped like the training data (see below).

Transformers (merged fp16)

from transformers import AutoModelForCausalLM, AutoTokenizer

tok = AutoTokenizer.from_pretrained("openbmb/MiniCPM5-1B")
model = AutoModelForCausalLM.from_pretrained(
    "RegeneratusLabs/augury-1b", subfolder="merged-fp16")

Input contract (this is how the model was trained)

System: You are AUGURY, a soil health assistant specializing in weeds and plants
        as soil indicators. You receive structured soil indicator data and present
        it in clear, conversational language suitable for farmers and land
        managers. ... Never invent or modify indicator data — only present what
        is provided. ... You do NOT provide management recommendations,
        herbicide advice, or agronomic prescriptions.

User:   Species: Capeweed (Arctotheca calendula)
        Region: Australia

        Indicators:
        - Moisture: dry to moderately dry, well-drained soils
        - Soil pH: neutral to slightly acidic
        - Fertility: moderate, tolerates low fertility

        What does this tell me about my soil?

In production the AUGURY funnel server builds this prompt from the species database (RegeneratusLabs/augury-species-db) — never from the model's memory.

Training


Base	openbmb/MiniCPM5-1B (1.1B)
Data	`RegeneratusLabs/augury-training-data` — 13,627 train / 1,514 val rows, AU-balanced (186/188 AU species)
Method	LoRA (r=16, alpha=32), 3 epochs, lr 2e-4 cosine, effective batch 32, bf16
Hardware	ModelScope PAI-DSW A10 24GB
Eval	val loss 0.0398 · local smoke: 6/6 fact-keys echoed, correct refusal

Limitations

AU-first: Australian species are fully represented; European Ellenberg rows dominate in raw count but AU species each carry region-tagged examples.
Fine-tuned on one format — deviate far from the input contract and quality drops.
Indicator claims are literature-based, not field-verified.

Files

MiniCPM5-1B-AUGURY-Q4_K_M.gguf — phone-ready (~660MB)
MiniCPM5-1B-AUGURY-F16.gguf — full precision GGUF
merged-fp16/ — merged safetensors for transformers
lora-adapter/ — the LoRA adapter (reproducibility)

License

Apache-2.0 (weights). Training data CC-BY-4.0 (see the dataset repos).

Project

Repo: github.com/RegeneratusLabs/AUGURY · Model card: 2026-08-12

Downloads last month: -

GGUF

Model size

1B params

Architecture

llama

Hardware compatibility

4-bit

16-bit

Model tree for RegeneratusLabs/augury-1b

Base model

openbmb/MiniCPM5-1B

Adapter

(55)

this model