GithubHelp home page GithubHelp logo

Comments (8)

ArtsiomWB avatar ArtsiomWB commented on September 21, 2024

Hi @YK-YoungK! Thank you for writing in.

Could you please send a link to your workspace where you are experiencing the first part of your graph not appearing in the UI?

from wandb.

YK-YoungK avatar YK-YoungK commented on September 21, 2024

Hi @YK-YoungK! Thank you for writing in.

Could you please send a link to your workspace where you are experiencing the first part of your graph not appearing in the UI?

Hi, here is the link: https://wandb.ai/yk-youngk/grok-mlp-default-2?workspace=user-yk-youngk.

from wandb.

ArtsiomWB avatar ArtsiomWB commented on September 21, 2024

Two quick follow-up questions:

  • could you please point me to the graph where you are seeing this concern? not having much luck locating it.
  • Are you logging data every single wandb step?

from wandb.

YK-YoungK avatar YK-YoungK commented on September 21, 2024

Two quick follow-up questions:

  • could you please point me to the graph where you are seeing this concern? not having much luck locating it.
  • Are you logging data every single wandb step?

I see this concern in the figure train_accuracy, val_accuracy in project train_pct=70. Besides, if the epoch x is logged, then the next logging epoch is max(1.01x, x+1).

from wandb.

ArtsiomWB avatar ArtsiomWB commented on September 21, 2024

Could you please add that figure to a report and then share it with me? Still no luck on my side.

I think what could be happening is our Sampling behavior. After we have logged >1500 points our chart sampling kicks in for performance reasons. If more than 1500 points are being displayed at once, some of the points get combined to prevent slowing down the UI. To confirm this what you can do is simply zoom in to the part of the graph that is missing in your UI and see if those points start appearing again.

from wandb.

ArtsiomWB avatar ArtsiomWB commented on September 21, 2024

Hi there, I wanted to follow up on this request. Please let us know if we can be of further assistance or if your issue has been resolved.

from wandb.

YK-YoungK avatar YK-YoungK commented on September 21, 2024

Could you please add that figure to a report and then share it with me? Still no luck on my side.

I think what could be happening is our Sampling behavior. After we have logged >1500 points our chart sampling kicks in for performance reasons. If more than 1500 points are being displayed at once, some of the points get combined to prevent slowing down the UI. To confirm this what you can do is simply zoom in to the part of the graph that is missing in your UI and see if those points start appearing again.

Hi, here is one report: https://api.wandb.ai/links/yk-youngk/0s97madx. Besides, when I zoom into the part of the missing curve the data seems to appear. Sorry for late response.

from wandb.

ArtsiomWB avatar ArtsiomWB commented on September 21, 2024

Thank you so much for sending it over! It does look like sampling. Our engineers are aware of the issue and are looking into a fix. Thank you so much for bringing this up!

from wandb.

Related Issues (20)

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.