DantinoX Logo

Get Started

  • Quickstart
  • Cookbook
  • Interactive Notebooks
  • DantinoX vs HuggingFace Transformers

Models

  • Architecture
  • Core Layers — Deep Reference
  • Paradigm System
  • Generation & Retrieval Paradigms
  • Autoregressive (AR) Generation
  • Discrete Diffusion
  • Fast-dLLM Block-wise DualCache
  • Confidence-Aware Parallel Decoding
  • Autoregressive vs Masked Diffusion — Head-to-Head
  • Retriever — Sentence Embeddings for RAG

Training

  • Training Guide
  • Autoregressive Training
  • Diffusion Model Training
  • Continuous Flow-Matching Training
  • Embedder Training
  • Optimizers & Schedules
  • Hyperparameter Sweeps
  • Multi-GPU Training: Data and Tensor Parallelism
  • Full Training Suite

Inference

  • Inference & Generation
  • Autoregressive Generation
  • Diffusion Generation
  • Continuous Flow-Matching Generation
  • KV-Cache Mechanics

Tutorials

  • Tutorials
  • Training Your First Model
  • Masked Diffusion Language Model
  • LoRA Fine-Tuning
  • RAG with DantinoX Embedders
  • Tutorial: Benchmarking a Model
  • Pushing to HuggingFace Hub

Benchmarks

  • Benchmarks
  • Ablation Studies
  • Experiments & Results
  • Profiling & Benchmarking Architecture

API Reference

  • API Reference
  • dantinox — Top-Level API
  • dantinox.paradigms
  • dantinox.trainer — Training API Reference
  • dantinox.profiling
  • dantinox.benchmarking
  • dantinox.visualization

Developer Guide

  • Developer Guide
  • Adding a Core Layer
  • Custom Paradigm
  • Custom Benchmark Task
  • Custom Chart

Project

  • Configuration Reference
  • CLI Reference
  • Contributing
  • FAQ & Troubleshooting
  • Changelog
DantinoX
  • Search

© Copyright 2026, Marco Simoni.

Built with Sphinx using a theme provided by Read the Docs.