Vladimir Haltakov

Cool, thanks! I'll update the documentation with this guide.

You can encode the photo using `model.encode_image` (see how it is used exactly [here](https://github.com/haltakov/natural-language-image-search/blob/main/03-process-unsplash-dataset.ipynb)) and search using the embedding in the same way as if it comes from the text.

I think this is how CLIP works. I've observed similar behavior - you don't really know, when CLIP doesn't know :) I guess he reason is that CLIP was not...

I've observed the same problem. Here two Colab notebooks if somebody wants to test: - 1.7.0: https://colab.research.google.com/drive/1KcHMeI2N-FFthMEYip3A1ZNbccdRK4XR?usp=sharing - 1.7.1: https://colab.research.google.com/drive/1h3pyWaZ0gA4DYWoip49W1skBmkxZ_gDC?usp=sharing With 1.7.0 the results on the second run are very...

Thanks for the suggestion! Makes sense :)

Yeah, agree, this is a feature

What do you think about this workflow? If you like it, I can implement it and create a PR.

Yeah, exactly! OK, I'll try to find time to implement it tomorrow and will share the PR.

This is possible to implement in Python for sure. I just don't have to do that, sorry