poloclub
How to add images?
Is there a notebook that shows how to add images with something like their captions? For instance, I have a csv file that contains "image_name" and "captions." I want to display the images and have the captions be used for topic modeling.
I am looking for the same feature, got confused with image_url_prefix in generate_grid_dict.
Hi! Sorry for the late response. You can check out this tutorial notebook to add images: https://github.com/poloclub/wizmap/blob/main/example/diffusiondb-images.ipynb
Hello @xiaohk. I ran both the notebook you linked to and your DiffusionDB example (https://poloclub.github.io/wizmap/?dataset=diffusiondb) and am unable to see the images in both. I have images checked both under points and "contour."
I just tested https://poloclub.github.io/wizmap/?dataset=diffusiondb and I can see the images. For this one you need to wait until the image points are loaded (they are loaded after the text points).
The notebook will create a slightly different format: each data point includes a text and an image. The original diffusiondb wizmap has separate text-only data points and image-only data points.
You can see the new wizmap created by the notebook here: https://poloclub.github.io/wizmap/?dataURL=https%3A%2F%2Fhuggingface.co%2Fdatasets%2Fxiaohk%2Fembeddings%2Fresolve%2Fmain%2Fdiffusiondb%2Fdata.ndjson&gridURL=https%3A%2F%2Fhuggingface.co%2Fdatasets%2Fxiaohk%2Fembeddings%2Fresolve%2Fmain%2Fdiffusiondb%2Fgrid.json
Ah. I see. Thank you for the follow-up @xiaohk . I think it would be helpful to have a loading count that displays the total percentage of text and images loaded so far.
It may also be beneficial to load a small batch of images and text together and then load the next batch. This may be more than you have bandwidth for at the moment. Just wanted to put the suggestion out there. Really enjoy the software.