FarmifAI_1.3_GGUF

FarmifAI 1.3 (GGUF)

GGUF versions of FarmifAI 1.3, a small language model (fine-tuned from Qwen3.5-0.8B) that answers agricultural questions in Spanish from a context you provide. It is built for Colombian agriculture and powers FarmifAI, an offline assistant app for farmers. These files run on llama.cpp and compatible tools, including on mobile devices.

FarmifAI is not a general-purpose chatbot. It is trained to answer from the context it receives, so it should always be used together with a retrieval step.

Files

Prompt format

The model was trained with a fixed Spanish system prompt. Use it as is:

Eres un asistente agrícola. Responde únicamente con la información dentro de <knowledge>. Si la respuesta no está en el contexto, declara que no tienes información; no inventes datos.

Instrucciones de formato:
- En <reasoning>, analiza paso a paso el contexto frente a la ...

<<< PASTE THE REST OF THE EXACT SYSTEM PROMPT FROM THE DATASET HERE >>>

The user message contains the retrieved context followed by the question:

<knowledge>
{retrieved context}
</knowledge>

{question}

The model replies with a step-by-step analysis and a final answer:

<reasoning>
Step-by-step analysis of the context against the question.
</reasoning>
<answer>
Final answer in Spanish, based only on the provided context.
</answer>

Applications typically show only the <answer> block. Parse it defensively in case the tags are missing.

Recommended settings: temperature=0.1, top_p=0.9, max 512 new tokens.

Quickstart

llama.cpp

# --no-mmproj skips the vision projector, which is not needed for text use
llama-server -hf FarmifAI/FarmifAI_1.3_GGUF:Q4_K_M --jinja --no-mmproj

import re
from openai import OpenAI

client = OpenAI(base_url="http://localhost:8080/v1", api_key="none")

SYSTEM_PROMPT = "..."  # the system prompt from "Prompt format" above

context = "..."  # passages retrieved from your knowledge base
question = "¿Cómo puedo controlar la broca en mi cultivo de café?"
user_message = f"<knowledge>\n{context}\n</knowledge>\n\n{question}"

response = client.chat.completions.create(
    model="FarmifAI_1.3",
    messages=[
        {"role": "system", "content": SYSTEM_PROMPT},
        {"role": "user", "content": user_message},
    ],
    temperature=0.1,
    top_p=0.9,
    max_tokens=512,
)

text = response.choices[0].message.content
match = re.search(r"<answer>(.*?)</answer>", text, re.DOTALL)
print(match.group(1).strip() if match else text)

Ollama, LM Studio and others

ollama run hf.co/FarmifAI/FarmifAI_1.3_GGUF:Q4_K_M

In these apps, set the system prompt and the recommended settings manually; without the system prompt the model will not follow the expected output format.

Limitations

• The model can still make mistakes or add details that are not in the context. Check its recommendations, especially anything about agrochemicals, doses or safety periods, and do not treat it as a substitute for professional advice.
• Answer quality depends on the quality of the retrieved context.

Links

• Full-precision model, training details and results: FarmifAI/FarmifAI_1.3
• Dataset: FarmifAI/FarmifAI_dataset_1.3

Fine-tuned and converted to GGUF with Unsloth.