Stuff

[gateway]

Btw thank you so much for taking the time to do this. I have been playing around with the prior code. I'm surprised I got this to run on windows as well (my Titan RTX is on windows).

Would you mind enabling the discussion tab in github for this?

Oh I did notice one issue on windows.. the output filename blows if you have things like ^ or | type of separators in your text prompt.

One question about video.. can you add a way to set a downscale size.. I have a bunch of 4k footage I wanted to test this on a downscale option when extracting the images from the video would be awesome.

I think the VF option in ffmpeg works.. -vf scale=192:168

Thoughts?

[rkhamilton]

Thanks for the note! I've just put this all together in the last week. I'll enable discussions shortly. I've never had anybody want to talk about any of my code before so I've never even looked at is. I assumed it was enabled! I also noticed that the filename breaks if you have phrase separators. In my examples I just manually enter filenames instead of sanitizing them. The simplest use case was just manual filenames, but I was always wanting to rename them to the text prompt. I'll think about whether I should break out some kind of "first safe part of the text prompt" method for naming files, or just rely on the user to use safe filenames manually. I want to warn you that I thought I was done with it and called it v 1.0, but a friend of mine suggested I incorporate ESRGAN upscaling, which excited me and I'm close to pushing a 1.1 that breaks compatibility with most stuff. My generate.*_video methods all wrapped up generating new images with encoding video to make for a compact structure, but that made it hard to easily integrate ESRGAN upscaling. The new syntax will look like generate video frames upscale video frames encode video. Re: downscaling video, I'll take a look. It should be easy enough. In your example is it just generating files that take up too much space, or is it failing to run?

…

On Sun, Oct 17, 2021 at 6:28 PM Gateway ***@***.***> wrote: Btw thank you so much for taking the time to do this. I have been playing around with the prior code. I'm surprised I got this to run on windows as well (my Titan RTX is on windows). Would you mind enabling the discussion tab in github for this? Oh I did notice one issue on windows.. the output filename blows if you have things like ^ or | type of separators in your text prompt. One question about video.. can you add a way to set a downscale size.. I have a bunch of 4k footage I wanted to test this on a downscale option when extracting the images from the video would be awesome. I think the VF option in ffmpeg works.. -vf scale=192:168 Thoughts? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#1>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACAEKSBY4AXFNGU2BS5PS63UHNEXZANCNFSM5GFJPNFQ> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

-- Ryan

[gateway]

I started a new discussion about this.. didn't realize this was moved to a discussion.

I had an idea though while thinking about you wanting to keep the text as a filename.. it would be great if we stored the text and settings in the metadata of the image. Exif data can be stored in the user-comment field. So if you wanted to go back to an image to see what settings you had you could just read the meta data.. thoughts? :)

[rkhamilton]

Ok, it's implemented for v1.2. Separate chunks for each relevant parameter.

[gateway]

Dude your awesome thank you so much for doing this an sometimes entertaining my crazy thoughts and ideas..

[rkhamilton]

Once you get a hold of 1.2 and generate your method for making annotated mosaics please share it with me. I was just scanning folders of thumbnail images to evaluate parameters.

[gateway]

sure thing.. this was done a while ago .. of course this was sooooo many different prompts from a semi random text generator. https://imgur.com/gallery/KIZuRyJ

[rkhamilton]

I was inspired by adding metadata to these images, so last night I worked up changes to add metadata to video as well.
title = text_prompts
comments = a string of generation parameter
example: iterations: 15, image_prompts: [], noise_prompts: [], init_weight_method: decay, init_weight 1.00, init_image video_frames, change_prompt_every 0, seed 796212828143400

In the next version, which I should push today once some automated tests finish.