9 repository-uri
Developer-defined logic for reading specific raw data samples from storage into tensors.
Distinct from Data Loading Optimizations: Focuses on the logic of reading training samples from disk, not database fetch optimizations.
Explore 9 awesome GitHub repositories matching data & databases · Custom Dataset Loading. Refine with filters or upvote what's useful.
This project is a comprehensive collection of educational examples and reference implementations for building vision and language models using PyTorch. It serves as a deep learning tutorial covering the end-to-end process of developing neural networks, from initial architecture definition to final production deployment. The repository provides detailed guides on implementing a wide range of domain-specific models, including convolutional neural networks for object detection and segmentation, as well as transformer and recurrent architectures for natural language processing. It emphasizes gene
Provides custom class implementations to define how individual data samples are read from storage.
mmcv is a foundation library for computer vision based on PyTorch. It provides a comprehensive system for constructing convolutional neural networks, a toolkit for image and video preprocessing, and a collection of high-performance deep learning vision operators. The project is distinguished by its hardware-accelerated kernels for complex operations such as deformable convolutions and region pooling. It features a configuration-driven framework that allows for the dynamic instantiation of network layers and the registration of custom modules without modifying code. The library covers a broad
Implements logic for loading raw image files and corresponding annotations from disk into tensors for training.
Snorkel is a weak supervision system that enables users to programmatically generate training labels for machine learning models without manual annotation. At its core, it provides a framework for writing labeling functions as Python callables that each vote on data points, and then trains a probabilistic graphical model over these multiple weak supervision sources to estimate latent true labels without any ground truth data. The system automatically learns accuracy and correlation parameters between labeling functions by analyzing observed agreement patterns on unlabeled data, converting lab
Provides a framework for writing Python callables that vote on data points to programmatically label datasets.
GDAL este o bibliotecă de traducere open source licențiată MIT care oferă un model de date abstract unificat pentru citirea și scrierea datelor geospațiale raster și vectoriale în sute de formate de fișiere. Servește ca o bibliotecă fundamentală de traducere a datelor geospațiale, permițând accesul la diverse formate de date geospațiale printr-o singură interfață consistentă. Biblioteca își expune funcționalitatea de bază prin utilitare de linie de comandă care permit utilizatorilor să traducă, convertească și proceseze date geospațiale între formate. Un motor de transformare a coordonatelor gestionează conversiile între sistemele de referință spațială, în timp ce un sistem de plugin-uri pentru drivere de format încarcă logica de citire și scriere specifică formatului la runtime. Stratul de sistem de fișiere virtual oferă acces uniform I/O prin fișiere locale, HTTP, stocare în cloud și arhive comprimate, iar un cache de blocuri raster gestionează cache-ul de tile-uri în memorie pentru a reduce operațiunile I/O. GDAL suportă citirea și scrierea datelor geospațiale atât raster, cât și vectoriale, cu iterarea caracteristicilor vectoriale care transmite caracteristicile individual fără a încărca seturi de date întregi în memorie. Proiectul permite interoperabilitatea geospațială între formate prin suportarea schimbului de date între diferite ecosisteme software geospațiale prin suportul său extins pentru formate.
Writes raster geospatial data to hundreds of file formats from a common in-memory model.
Anomalib is a PyTorch-based library for visual anomaly detection, offering a modular framework, a comprehensive model zoo, and a benchmarking suite designed for industrial defect detection. It provides a wide range of algorithms—including generative, discriminative, teacher-student, and vision-language approaches—that support unsupervised, few-shot, and zero-shot settings. The library enables deployment through model export to ONNX and OpenVINO for edge devices, and includes a no-code web application for training and inference. It also features a command-line interface for orchestrating multi
Anomalib loads image datasets from custom folders and standard formats, structuring them for model training and evaluation.
TorchGeo is a PyTorch library designed for deep learning on geospatial data, providing a framework for building and training neural networks for tasks such as semantic segmentation, object detection, and change detection. It serves as a comprehensive pipeline for remote sensing, featuring specialized dataset loaders and multispectral image preprocessing tools. The library is distinguished by a dedicated remote sensing model zoo and extensive support for transfer learning, allowing users to integrate pre-trained weights optimized for specific satellite sensors. It also includes support for sel
Implements specialized loading for raster files incorporating custom coordinate reference systems and band selection.
PyTorchZeroToAll este o resursă educațională și o colecție de tutoriale axate pe deep learning și framework-ul PyTorch. Oferă un parcurs de învățare structurat pentru implementarea arhitecturilor de rețele neuronale, variind de la sintaxa de bază a limbajului și fundamente, până la design-ul complex de modele. Proiectul servește ca ghid de implementare pentru construirea diverselor tipuri de rețele, inclusiv rețele liniare, logistice, convoluționale și recurente. Acoperă în mod specific fluxul de lucru pentru modelarea secvențelor prin utilizarea mecanismelor de atenție și a rețelelor la nivel de caracter. Resursa acoperă, de asemenea, pregătirea datelor pentru machine learning, inclusiv utilizarea încărcătoarelor de seturi de date personalizate, instrumente de procesare a textului și traducerea fișierelor sursă brute în tensori pentru pipeline-urile de antrenare.
Demonstrates how to implement custom dataset loading classes to translate raw files into PyTorch tensors.
mmpretrain is a modular PyTorch computer vision framework designed for developing, training, and benchmarking deep learning architectures. It serves as a comprehensive toolkit for vision tasks, providing a specialized platform for multimodal machine learning and self-supervised learning. The project features a computer vision model zoo containing architectural definitions and pre-trained weights for backbones such as ViT, ConvNeXt, and Swin Transformer. It distinguishes itself through a dedicated self-supervised learning toolkit that implements algorithms like MAE and DINO to train models wit
Implements logic for reading local images and labels from JSON, YAML, or Pickle formats into tensors.
This is a structured deep learning curriculum for programmers, delivered as a collection of Jupyter notebooks. It teaches the fundamentals of training neural networks for computer vision, natural language processing, tabular data analysis, and collaborative filtering using PyTorch and the fastai library. The course is designed to be hands-on, guiding learners from building a training loop from scratch to fine-tuning pretrained models for a variety of practical tasks. The curriculum distinguishes itself by covering the full lifecycle of a deep learning project, from data preparation and augmen
Ships utilities for extracting labels from image filenames using custom functions or regex.