search

daita-technologies / interface

React-based user interface for the DAITA platform.
http://app.daita.tech
GNU Affero General Public License v3.0
1 stars 0 forks source link

Open data set list #4

Open pcaversaccio opened 2 years ago

pcaversaccio commented 2 years ago

We should have a list or index (i.e. links) of open data sets for computer vision data that we recommend on the platform (as a separate page of course). Also, later on (as a pro feature) it could be desirable that the user could easily integrate a selected data set into the dataset without downloading and uploading it separately to his/her data set. -> we must make everything convenient for the user. What are your thoughts?

AI-Daita commented 2 years ago

Ah, yes, this is also my question. I agree with a list of example data sets existed on the system. But, we should clearify something here:

  1. Will we disable the function "upload data" in MVP? Users do not need to upload, but they can download the results, correct?
  2. Do you mean "integrate" is "merge" (existing) dataset?
pcaversaccio commented 2 years ago

Re 1): Ok, let me clarify: the user must be able to upload his/her data of course and also download the results accordingly. What I mean by list/index of open data sets is just a page where the user can inform himself about what is generally available as open-source data set and can scroll through it and click on the reference link if his interested. This is more for informational purposes and is not related to the sandbox data set we provide. Like that: image

Re 2): yes, imagine you have your own data set uploaded and want to enrich it automatically with an open dataset that is available out there and the user thinks provides a value-add to his/her model. Instead of separating downloading it and adding it to an existing data set he/she uses on our platform, the user can just simply integrate (=merge) this data set in our platform with one click.

AI-Daita commented 2 years ago

Re 1): If I well understand, we have two different things: 1. Sandbox dataset is available in the system (with all files). 2. open dataset is just a page with information about available online dataset, If they want to use it, follow the liink and download the data then upload to our system. Is it corrrect? R2 2) Ok, got it.

pcaversaccio commented 2 years ago

Re 1): If I well understand, we have two different things: 1. Sandbox dataset is available in the system (with all files). 2. open dataset is just a page with information about available online dataset, If they want to use it, follow the liink and download the data then upload to our system. Is it corrrect?

Yes, this is correct. As some sort of inspiration, let me share this one: https://scale.com/open-datasets

pcaversaccio commented 2 years ago

Slightly related to this roadmap item: https://github.com/daita-technologies/roadmap/issues/21