Modelpublic

Raidone/almaaniraid-modelfiles

sourceHugging Facemitupdated 2mo agoView on Hugging Face

0likes

Model Card

<h1 align="center">Almaaniraid Modelfiles</h1>

<p align="center">Italian-language Ollama Modelfiles for professional AI assistants — built on <strong>Qwen2 7B</strong>.</p>

Description

These are Ollama Modelfiles that specialize the base qwen2:7b model into 15 professional AI assistants. Each Modelfile defines a SYSTEM prompt and inference parameters (temperature, top_p, context length) tuned for a specific domain.

All models respond in Italian by default.

Base Model

All Modelfiles use qwen2:7b as the base model:

bash

ollama pull qwen2:7b

Usage

bash

# 1. Pull the base model
ollama pull qwen2:7b

# 2. Create a specialized model from a Modelfile
ollama create Rai-ArchV1 -f Modelfile.architetto

# 3. Run it
ollama run Rai-ArchV1

Alternatively, download the Modelfile directly and run:

bash

ollama create Rai-ArchV1 -f /path/to/Modelfile.architetto

Available Models

Model Name	Modelfile	Role	Description
Rai-AiV1	`Modelfile.raiai`	Chief AI Coordinator	Orchestrator & general-purpose assistant
Rai-AxV1	`Modelfile.raiax`	Rapid Action Executor	Scripting, automation, DevOps
Rai-OpV1	`Modelfile.raiops`	Operations & Security	Sysadmin, security, infrastructure
Rai-KaiV1	`Modelfile.raikai`	AI/Data Scientist	ML, analytics, AI strategy
Rai-MtV1	`Modelfile.mythos-rdt`	Narrative Intelligence	Documentation, storytelling, writing
Rai-ArchV1	`Modelfile.architetto`	Architect / Civil Engineer	Building design, structural engineering, urban planning
Rai-AvvV1	`Modelfile.avvocato`	International Lawyer	International law, contracts, compliance
Rai-ChiV1	`Modelfile.chimico`	Chemist / Professor	Organic/inorganic chemistry, lab research
Rai-ComV1	`Modelfile.commercialista`	Accountant	Tax, accounting, corporate advisory
Rai-DtnV1	`Modelfile.dottore`	Doctor / Biologist / Dietitian	Clinical nutrition, diagnostics, biology
Rai-FarV1	`Modelfile.farmacista`	Pharmacist	Pharmacology, drug interactions, herbal medicine
Rai-SciV1	`Modelfile.scienziato`	Cosmic Scientist	Astrophysics, cosmology, quantum physics
Rai-VetV1	`Modelfile.veterinario`	Veterinary Dietitian	Pet nutrition (dogs & cats)
Rai-MecV1	`Modelfile.meccanico`	Mechanical Technician	Mechatronics, auto repair, diagnostics
Rai-LLMV1	`Modelfile.llm_expert`	LLM Engineer	AI/ML training, fine-tuning, deployment

Parameters

Each Modelfile sets optimized inference parameters:

Parameter	Range	Purpose
`temperature`	0.2–0.8	Precision (low) vs creativity (high)
`top_p`	0.8–0.92	Nucleus sampling diversity
`num_ctx`	8192–16384	Context window length (tokens)

License

MIT