ls-mlkit Documentation

Welcome to ls-mlkit, a comprehensive machine learning toolkit that provides various utilities for deep learning, diffusion models, optimization, and data processing.

Installation

Installation:

• Installation Guide

API

API Reference:

• ls_mlkit
  • ls_mlkit package
    • Subpackages
      • ls_mlkit.dataset package
      • ls_mlkit.diffuser package
      • ls_mlkit.model package
      • ls_mlkit.optimizer package
      • ls_mlkit.pipeline package
      • ls_mlkit.scheduler package
      • ls_mlkit.util package
    • Module contents

Features Overview

Pipeline BasePipeline, DistributedPipeline, MyDistributedPipeline

Diffuser DDPM, DDIM, and SO(3)VPSDE

Scheduler cosine(_with_warmup), linear(_with_warmup), constant, cosine_annealing

Optimizers KFA, SAM

Flow Matching EuclideanFlow

Datasets

Numerical datasets:

IrisDataset

Image datasets:

MNIST, FashionMNIST, CIFAR10, CIFAR100

Language datasets:

LDADataset, MT19937, RegularLanguageDataset, meta-math/MetaMathQA, gsm8k, glue/sst2, m-a-p/CodeFeedback-Filtered-Instruction, silk-road/Wizard-LM-Chinese-instruct-evol, tatsu-lab/alpaca

Models LongLinearModel, CausalLanguageModelForAuto

Utils Various tools

Indices and tables

• Index

• Module Index

• Search Page