machinelearning-samples
machinelearning-samples copied to clipboard
Big files excluded from: ObjectDetection sample (C# and F#)
Hi all!, here is my journey for a few months: 1°) Datasets (and graphs) are too big to be included in the github repository (more than 2 Gb!) 2°) There are already examples using dataset download in the ML samples (but some of them fail) 3°) Some datasets are common to C# and F# (and soon VB!) 4°) So I will host datasets on my personal website 5°) To avoid putting my name on URLs, I will use url redirection (like https://bit.ly/) 6°) Net Core does not follow redirection from https to http, for new, higher security reasons 7°) However, it is still possible to follow the redirection explicitly, this is finally reasonable behavior 8°) But with .Net6, WebClient is now obsolete! 9°) Downloading a file is now very easy with .Net6... using Flurl.Http NuGet library! (we cannot find any official example to replace WebClient by httpClient as it is yet recommended)
Should I do something, or just wait? I can replace Flurl.Http with WebClient, but WebClient is now deprecated with .Net6.