Examples: query, "exact match", wildcard*, wild?ard, wild*rd
Fuzzy search: cake~ (finds cakes, bake)
Term boost: "red velvet"^4, chocolate^2
Field grouping: tags:(+work -"fun-stuff")
Escaping: Escape characters +-&|!(){}[]^"~*?:\ with \, e.g. \+
Range search: properties.timestamp:[1587729413488 TO *] (inclusive), properties.title:{A TO Z}(excluding A and Z)
Combinations: chocolate AND vanilla, chocolate OR vanilla, (chocolate OR vanilla) NOT "vanilla pudding"
Field search: properties.title:"The Title" AND text
Answered
Thank You All For Taking The Time To Answer Our Survey (If You Haven'T Already, We Urge You To

Thank you all for taking the time to answer our survey (If you haven't already, we urge you to do so ).
Your feedback has a major impact on what we build, don't be shy :)

Here are a few features that you requested that we will have in the upcoming 0.15 release (due at the end of this month).

  • Column resizing / ordering / filtering
  • The much anticipated Tags :)
  • Automation / Parameter Sweeping / HyperBand Bayesian Optimization
  • Custom Task Types
  • Trains-Agent Services mode (i.e launch multiple tasks on a cpu node)
  • Plug-able Cloud Auto Scaler (AWS and more)
  • Improved Audio Support
    Things we are actively working on:
  • Controllers
  • Nested projects
  • Configurable Alerts
    A few things we  already have  that might have gone unnoticed:
  • The introduction of trains.StorageManager for any url download/upload with built in cache (http/s3/gs/azure support)
  • All artifacts are now cached including auto unpacking of zipped folders
  • Reporting Pandas tables
  • Reporting histograms (stacked/grouped)
  • Searchable comment/description (See "Info" tab, "Description" section)
    As always, if you feel something is missing or looking for the best way to implement a solution, ask away, don't be shy.

May the force be with you 🚆

  
  
Posted 4 years ago
Votes Newest

Answers 7


Hi @<1523722267119325184:profile|PunySquid88> I guess it's a good thing we talk, because I believe that what you are looking for is already available :)

Logger.current_logger().report_media('title', 'series', iteration=1337, local_path='/tmp/bunny.mp4')

This will actually work on any file, that said, the UI might display the wrong icon (which will be fixed in the next version).

We usually think of artifacts as data you want to reuse, so all the files uploaded there are accessible from anywhere through the Task object. Logging is more for debug purposes, and actually can store multiple copies of the same file, with diff iteration numbers.

Here's an example from the demoapp server
None

  
  
Posted 4 years ago

one thing I don't see in the list of feature requests that would be nice would be support for comparing videos 🙂 . Right now I'm able to upload video output as an artifact and view it from the website, but I would like to be able to see the videos in a grid when comparing various experiments

  
  
Posted 4 years ago

I think we added it somewhere in 0.14, anyhow I just checked the Logger doc, it is there now 🙂

  
  
Posted 4 years ago

image

  
  
Posted 4 years ago

for reference, this is what it looks like in Weight and Biases

  
  
Posted 4 years ago

omg amazing

  
  
Posted 4 years ago

figured it was only for images, as shown in the documentation

  
  
Posted 4 years ago
510 Views
7 Answers
4 years ago
8 months ago
Tags