AIMLite Studio — Zero-Path AI & ML Project Creator

1

Project Identity

Project Name (default: my_ai)

Target Directory (default: .)

2

Select System Paradigm Template

Scratch / Custom ML

Tabular, supervised, deep learning & classification models.

Scikit-learn • PyTorch

RAG Knowledge Base

Semantic vector search, grounded LLM synthesis & HyDE.

Vector DB • Smart Chunk

LoRA Adapter Tuning

Low-rank parameter adaptation & dynamic hot swapping.

PEFT • Hugging Face

3

RAG Architecture Configurations

Chat Provider for LLM Synthesis

OpenAI

gpt-4o-mini

Google Gemini

gemini-1.5-flash

Anthropic

claude-3-5

Ollama Local

llama3.2

Local Pipeline

meta-llama

Mock Provider

Zero external API

Model Identifier

Vector Database Backend

In-Memory (JSON)

Zero external database

PostgreSQL + pgvector

Enterprise SQL ORM

Embedding Engine

SentenceTransformers

all-MiniLM-L6-v2

Provider API Native

Native API embeddings

Hash TF-IDF

Fast zero-dependency

3

LoRA Fine-Tuning Hyperparameters

Rank r

Alpha α

Base Model

4

Code Generation Mode (--clean)

Clean Project Skeleton (--clean)

Generates pristine 0-byte code files with abstract class contracts and zero sample data.

5

Add Extra Packages (pip)

Quick add:

6

Environment & Dependency Setup

Create Isolated .venv

Provisions a managed local virtual environment inside the project directory.

Install Dependencies Now (--install)

Immediately runs pip install into .venv (Vite-style: deferred by default).