Location of images saved changed for single images output/output?

[Gateway]

This block of code

vqgan_clip.generate.single_image(eng_config = config,
        text_prompts = text_prompts,
        iterations = 50,
        save_every = 10,
        output_filename = 'output' + os.sep + 'man1')

Saves in a folder output/output/man1.png

this block of code

vqgan_clip.generate.single_image(eng_config = config,
        text_prompts = text_prompts,
        iterations = 50,
        save_every = 10,
        output_filename = 'man1')

saves it in the root folder.. not output folder..

[Gateway]

I think this is throwing off all other functions like video generation with init images etc.. Did something change in the way output is handled ?

[Ryan Hamilton]

This is a bug related to 24. The same sanitization function had a folder bug. It will be fixed asap.

[Ryan Hamilton]

The expected behavior is what you describe above.

output_filename	save location
'output' + os.sep + 'man1'	.\output\man1.png
'man1'	.\man1.png
'man1.png'	.\man1.png
'.\man1.jpg'	.\man1.png

I want to move to a syntax of specifying output file name extensions, but warn if the user requests anything other than png. Output filenames with non-png extensions were being rewritten to have the correct extension. There was a bug in this code v1.2.0, but it should be fixed now. Let me know if you see unexpected behavior.

[Gateway]

should I be pulling in tags via pip somehow or should i be pulling from main?

[Ryan Hamilton]

I assumed people will just use pip to install like in the readme installation instructions, rather than cloning the repo. In that case, you would just pip upgrade.

pip install git+https://github.com/rkhamilton/vqgan-clip-generator.git --upgrade

If you are cloning the repo that’s fine, of course.