Posts
-
swin-transformer-pytorch
Implementation of the Swin Transformer in PyTorch.
Implementation of the Swin Transformer architecture. This paper presents a new vision Transformer, called Swin Transformer, that capably serves as a general-purpose backbone for computer vision. Challenges in adapting Transformer from language to vision arise from differences between the two domains, such as large variations in the scale of visual entities and the high resolution of pixels in images compared to words in text. To address these differences, we propose a hierarchical Transformer whose representation is computed with shifted windows. The shifted windowing scheme brings greater efficiency by limiting self-attention computation to non-overlapping local windows while also allowing for cross-window connection. This hierarchical architecture has the flexibility to model at various scales and has linear computational complexity with respect to image size. These qualities of Swin Transformer make it compatible with a broad range of vision tasks, including image classification (86.4 top-1 accuracy on ImageNet-1K) and dense prediction tasks such as object detection (58.7 box AP and 51.1 mask AP on COCO test-dev) and semantic segmentation (53.5 mIoU on ADE20K val). Its performance surpasses the previous state-of-the-art by a large margin of +2.7 box AP and +2.6 mask AP on COCO, and +3.2 mIoU on ADE20K, demonstrating the potential of Transformer-based models as vision backbones.
This is NOT the official repository of the Swin Transformer. At the moment in time the official code of the authors is not available yet but can be found later at: https://github.com/microsoft/Swin-Transformer.
All credits go to the authors Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin and Baining Guo.
Tags: #python • deep-learning • machine-learning
-
nanoid.rb
Ruby implementation of Nanoid, secure URL-friendly unique ID generator
This is ReadMe for version 2.0, if you looking for version compatible with NanoID 1.x look in branch v1.x.
A tiny, secure URL-friendly unique string ID generator.
Ruby implementation of original NanoID https://github.com/ai/nanoid
Tags: #ruby
-
offlineimap
Read/sync your IMAP mailboxes (python2) [LEGACY: move to offlineimap3]
“Get the emails where you need them.”
IMPORTANT NOTE: This repository is for python2 only. The support for offlineimap3 is happening in [Official offlineimap for python3][offlineimap3].
I’ll still lazily maintain this legacy offlineimap but users should definitely go with offlineimap3.
- [Official offlineimap for python3][offlineimap3].
- [Official offlineimap for python2][offlineimap].
Tags: #python • imap • offline
-
EnchantmentCracker
Cracking the XP seed in Minecraft and choosing your enchantments
Cracking the XP seed in Minecraft and choosing your enchantments.
Tags: #java
-
jsondiff
Diff JSON and JSON-like structures in Python
Diff JSON and JSON-like structures in Python.
Tags: #python • json • diff
-
resque-scheduler
A light-weight job scheduling system built on top of Resque
Resque-scheduler is an extension to Resque that adds support for queueing items in the future.
Job scheduling is supported in two different ways: Recurring (scheduled) and Delayed.
Scheduled jobs are like cron jobs, recurring on a regular basis. Delayed jobs are resque jobs that you want to run at some point in the future. The syntax is pretty explanatory:
Resque.enqueue_in(5.days, SendFollowupEmail, argument) # runs a job in 5 days, calling SendFollowupEmail.perform(argument) # or Resque.enqueue_at(5.days.from_now, SomeJob, argument) # runs a job at a specific time, calling SomeJob.perform(argument)Tags: #ruby • scheduler • queue
-
rake-compiler
Provide a standard and simplified way to build and package Ruby C and Java extensions using Rake as glue.
rake-compiler is first and foremost a productivity tool for Ruby developers. Its goal is to make the busy developer’s life easier by simplifying the building and packaging of Ruby extensions by simplifying code and reducing duplication.
It follows convention over configuration by advocating a standardized build and package structure for both C and Java based RubyGems.
rake-compiler is the result of many hard-won experiences dealing with several diverse RubyGems that provided native extensions for different platforms and different user configurations in different ways. Details such as differences in code portability, differences in code clarity, and differences in project directory structure often made it very difficult for newcomers to those RubyGems.
From these challenges, rake-compiler was born with the single-minded goal of making the busy RubyGem developer’s life much less difficult.
Tags: #ruby
-
Services_Openstreetmap
Makes communicating with the Open Street Map API, and Nominatim, from PHP intuitive.
OpenStreetMap is a global project with an aim of collaboratively collecting mapdata. This package aims to make communicating with the OSM API intuitive.
Tags: #php • services-openstreetmap • mapquest-api
-
gazpacho
🥫 The simple, fast, and modern web scraping library
gazpacho is a simple, fast, and modern web scraping library. The library is stable, and installed with zero dependencies.
Tags: #python • gazpacho • webscraping
-
optica
A tool for keeping track of nodes in your infrastructure
Optica is a service for registering and locating nodes. It provides a simple REST API.
Nodes can POST to / to register themselves with some parameters. Humans can GET / to get a list of all registered nodes. GET also accepts some parameters to limit which of the registered nodes you see.
Tags: #ruby
subscribe via RSS