Library of Congress reporthttps://labs.loc.gov/static/labs/work/reports/Cordell-LOC-ML-report.pdf?loclr=blogsig

So you’re almost creating portals!

Will you be sharing your slides? I know there will be a recording, but can we have the slides, too?

I’’d be really interested in whether you have user feedback or other data about how the images that get viewed in something like Photogrammar differ from which ones get seen more in other UIs

(If that’’s a question that you asked about this)

(Sorry for the weird double quotes; my keyboard is haunted)

Thanks! That’s very cool

Photogrammar.org

Google Colab notebookhttps://colab.research.google.com/drive/1JEPMi1uMlN3FSPB7lu-CSqfT7FT6WOVZ?usp=sharing

http://bit.ly/images-harvard2021

How do you decide which terms to use to describe things & stuff?

Is there a controlled vocabulary?

Please, if you have any questions on getting the Colab notebook open, you can ask for help in the chat

it is requiring a google login

ERROR: albumentations 0.1.12 has requirement imgaug<0.2.7,>=0.2.5, but you'll have imgaug 0.2.9 which is incompatible.

Shout out to Jeff Steward for agreeing to make this data available!

One question on this bit: To do this, click on the menu above and select the option Runtime > Restart runtime.

Not sure about which menu is meant.

OK, will wait!

ModuleNotFoundError: No module named 'dvt_workshop'

Never mind.

How is everybody? Good?

what does the GPU runtime message mean?

all good!

thanks!

My brain is exploding from so much useful/fascinating information at once

+1 x 12!

Great to hear!

This takes time - be patient

I get the answer: 4 person, 1 dog,2 ties…..

Yes

But that it found the ties is so funny

559

My image (125) broke up a single baby into 3 people

I’ve been surprised, when running the demos on your web site, how often it finds (or thinks it finds) ties. It loves #menswear!

Yes, mine has ties, too!

Sunday best for the portrait photographers

Exactly!

Have yet to see it find a bow tie or pocket square, but I’m holding out hope

I wonder what it would do with our virtual backgrounds on Zoom!!!

does Harvard have any plans to integrate tools like this into our search platforms?

Mine seems to be stuck at 975 of 1000. Can I run the next step anyway?

Oh, good — thank you!

If the little icon on the top left of the cell is no longer spiraling, that means the cell is no longer running

That is very helpful - thanks to all!

Is any part of this bandwidth or local machine resource dependent? Mine seemed to run a lot faster.

I am contemplating an 18th century Gothic called Spinning Wheel of Death.

I’ll be quiet!!!

Love it!

And bad jokes.

That makes me wonder about associating date ranges with terms/categories.

I love this but I have another meeting. Will you share the recording? Thank you for this workshop!

Yes, we will.

Thank you for coming!

Thank you very much!

Is there an easy way to see the list of potential objects that this script is set up to detect? (Apologies if you already covered this)

Thanks!

550 gives a creepy collection.

999 gives small children

Oh wow, 550

You’re not kidding, Erinlea!

Babies and coffins

Vertical wall textures?

It looks like corrupt images come up in the set for 600.

967 was much more banal!

336 is an architectural model, which is of interest for architecture people

Disappointingly, 666 is just palm trees and vehicles

Some people think palm trees and vehicles might be exactly apt for 666.

It’s python so shouldn’t we be checking 665? It looks much more appropriate :)

LOL

Fair point, Jess!

665… love it!

I found one image of… questionable costuming

(But I’m not finding anything similar to that image, thankfully…?)

I really appreciate the clear, thoughtsu pressntation that we could all try out!

https://harvard.az1.qualtrics.com/jfe/form/SV_b2cTP0zwv3xqaSG

https://github.com/amenglong/art_generation_cnn

thoughtful presentation, I meant!

The list of objects!

Thanks!

This was an excellent workshop! Thank you so much, Carol, Lauren, and Taylor for all the work you put into this to make it understandable. And the technology all worked!! Brava!

Qualtrics survey here: https://harvard.az1.qualtrics.com/jfe/form/SV_b2cTP0zwv3xqaSG

https://distant-viewing.github.io/dvt/index.html

Is the list in that git repo? Sorry to be dim.

https://github.com/nightrome/cocostuff/blob/master/labels.md

This was awesome. I fully agree with Rhea. Thank you!

Excellent workshop

Thank you so much!

Thanks, that was fantastic!

Thanks so much, this was really enlightening!

Thank you so much! This was excellent, much appreciated!

Thanks!

Thanks so much. This has been fantastic!

Thanks so much for putting this together!

This was lovely

This was fun and riveting!

Thank you!

Thank y’all so much!!

It was quite thrilling!

Great stuff, thanks a million