AIMLite Studio v2.2.2 — Visual AI & ML Project Scaffolder

1

Project Identity

package: my_ai

Project Name identifier

Target Directory command directory

                location: .
              

2

Architecture Paradigm

canonical template

Scratch / ML

✓

Supervised classification, regression, PyTorch & Scikit-Learn workflows.

Tabular • Neural

RAG Engine

✓

Enterprise knowledge QA, vector retrieval, smart chunking & HyDE synthesis.

Vector DB • Citations

LoRA Adapter

✓

Parameter-efficient fine-tuning, rank decomposition & live weight merging.

PEFT • Low-Rank

3

RAG Engine Parameters

retrieval & synthesis

Chat Provider

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 Store Backend

In-Memory (JSON)

Zero database setup required

PostgreSQL + pgvector

Enterprise vector persistence

Embedding Engine

SentenceTransformers

all-MiniLM-L6-v2

Provider API Native

Embedded via chat provider

Hash TF-IDF

Lightweight zero-dependency

3

LoRA Hyperparameters

rank & target modules

Rank r

Alpha α

Base Foundation Model

4

Generation Mode

boilerplate vs skeleton

Clean Skeleton Mode (--clean)

Generates 0-byte code interfaces with class contracts and zero dummy data.

5

Dependencies

empty on start — add packages as needed

Quick add:

6

Environment & Execution

runtime provisioning

Create Isolated .venv

Provisions a managed Python virtual environment inside the project directory.

Install Dependencies Now (--install)

Installs packages immediately during project creation (deferred by default).

Scaffold AI Application