Instructions to use edoggy/commodities-crop-analyst-7b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use edoggy/commodities-crop-analyst-7b with PEFT:

from peft import PeftModel
from transformers import AutoModelForCausalLM

base_model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen2.5-7B-Instruct")
model = PeftModel.from_pretrained(base_model, "edoggy/commodities-crop-analyst-7b")

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use edoggy/commodities-crop-analyst-7b 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 edoggy/commodities-crop-analyst-7b:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf edoggy/commodities-crop-analyst-7b:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf edoggy/commodities-crop-analyst-7b:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf edoggy/commodities-crop-analyst-7b:Q4_K_M

Build from source code

git clone https://github.com/ggml-org/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-app
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama serve -hf edoggy/commodities-crop-analyst-7b:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama cli -hf edoggy/commodities-crop-analyst-7b:Q4_K_M

Use Docker

docker model run hf.co/edoggy/commodities-crop-analyst-7b:Q4_K_M

Llama new
LM Studio
Jan

vLLM

How to use edoggy/commodities-crop-analyst-7b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "edoggy/commodities-crop-analyst-7b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "edoggy/commodities-crop-analyst-7b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/edoggy/commodities-crop-analyst-7b:Q4_K_M

Ollama
How to use edoggy/commodities-crop-analyst-7b with Ollama:
```
ollama run hf.co/edoggy/commodities-crop-analyst-7b:Q4_K_M
```
Unsloth Desktop

How to use edoggy/commodities-crop-analyst-7b with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf edoggy/commodities-crop-analyst-7b: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": "edoggy/commodities-crop-analyst-7b:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use edoggy/commodities-crop-analyst-7b with Docker Model Runner:
```
docker model run hf.co/edoggy/commodities-crop-analyst-7b:Q4_K_M
```

Lemonade

How to use edoggy/commodities-crop-analyst-7b with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull edoggy/commodities-crop-analyst-7b:Q4_K_M

Run and chat with the model

lemonade run user.commodities-crop-analyst-7b-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use edoggy/commodities-crop-analyst-7b with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf edoggy/commodities-crop-analyst-7b: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 edoggy/commodities-crop-analyst-7b:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use edoggy/commodities-crop-analyst-7b with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf edoggy/commodities-crop-analyst-7b: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 "edoggy/commodities-crop-analyst-7b: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"

commodities-crop-analyst-7b (v1)

A tool-calling LoRA on Qwen2.5-7B-Instruct that answers crop-supply questions only from data it queries — USDA CDL county acreage (11 states, 2018–2024), the 2022 Census of Agriculture, IBGE São Paulo yields, USDA report dates, and front-month futures for sugar, orange juice, corn, soybeans, wheat and cotton. It never states a number it didn't compute. Self-contained: the data, tools and runtime ship in this repo.

What it does

Area, change, rankings, state comparisons from the CDL; sugarcane model-vs-CDL-vs-census validation by county.
Price history and a statistical price projection (price_forecast): seasonal drift + volatility cone with an 80% interval and the method's backtested error (6–11% at 30 days; interval coverage 78–84%). The point estimate has no directional edge and the model says so. "Should I buy?" gets context, not a call.
Next-year Louisiana sugarcane expansion from sugarcane-forecast-v2 (new-cane recall 0.95 / precision 0.83; removals under-predicted).
Renders sugarcane probability maps for 20 LA/FL counties (sugarcane-classifier-v1) and price charts.
Asks a clarifying question when a year or place is missing; says so when something is outside the data.

Evaluation (785 held-out questions whose phrasings never appeared in training)

	Q4_K_M GGUF (Ollama)	LoRA adapter (4-bit HF)
Correct tool chosen	95.2%	95.6%
Exact arguments	92.6%	92.9%
Executed result identical to gold	92.6%	92.9%
Malformed tool calls	0	0
Underspecified / out-of-scope handled without a tool call	177 / 177	177 / 177

Run it (Ollama — any machine with ~6 GB free RAM/VRAM)

pip install gradio requests pandas pyarrow rasterio matplotlib joblib scikit-learn
hf download edoggy/commodities-crop-analyst-7b --local-dir commodities-crop-analyst && cd commodities-crop-analyst
ollama create commodities-crop-analyst -f Modelfile
python runtime.py                                  # chat UI at http://localhost:7860

Or with the LoRA adapter on a CUDA GPU (12 GB, 4-bit): pip install transformers peft bitsandbytes accelerate then python runtime.py --backend hf --adapter .

Each reply shows the answer, any image, and the exact tool call + result underneath so every number can be checked.

Training

12,000 synthetic examples generated from the bundled tables (45_gen_v2.py): questions sampled from 24–66 LLM-paraphrased phrasings per intent (44_paraphrase.py), gold tool calls, gold answers computed from the data, plus clarify/out-of-scope cases. QLoRA r=16, 400 steps, lr 1e-4, on one RTX 5070.

Limits — read before using

Not a trading tool. Satellite/CDL area for a year is only available after USDA's own reports; nothing here has a timing edge. Forecasts are baselines with error bars, not predictions of direction.
Long-horizon price projections (>120 days) are under-calibrated in backtests for volatile contracts (e.g. orange juice); the tool flags this in its output.
US coverage only (LA, FL, TX, IA, IL, IN, MN, NE, KS, ND, GA), 2018–2024; one Brazilian state's yields; no stocks, demand, FX or in-season weather.
Acreage is map-derived; county year-over-year changes under ~5% are within mapping noise.
A 7% tool-argument error rate means: check the tool call shown under each answer for anything that matters.

Base model: Qwen2.5-7B-Instruct (Apache-2.0). Data: USDA NASS (public domain), IBGE, NASA POWER, Yahoo Finance, Google Satellite Embedding V1 (CC-BY-4.0) via the linked models.

Downloads last month: 67

GGUF

Model size

8B params

Architecture

qwen2

Hardware compatibility

4-bit

Model tree for edoggy/commodities-crop-analyst-7b

Base model

Qwen/Qwen2.5-7B

Finetuned

Qwen/Qwen2.5-7B-Instruct

Adapter

(2,909)

this model

edoggy
/

commodities-crop-analyst-7b

commodities-crop-analyst-7b (v1)

What it does

Evaluation (785 held-out questions whose phrasings never appeared in training)

Run it (Ollama — any machine with ~6 GB free RAM/VRAM)

Training

Limits — read before using

Model tree for edoggy/commodities-crop-analyst-7b

Datasets used to train edoggy/commodities-crop-analyst-7b