Ryan Ramos

I'm not necessarily pushing for it, but I will say that what methods like Prompt-to-Prompt and EDICT have over pix2pix zero is the lack of a need to generate source...

Hi @Luke2642 , actually that's where I stopped too. I was able to get the code to run without runtime errors but always ended with black images (also ended up...

@Luke2642 Sorry for the late reply: got the exact same thing you did ("original and vqvae image fine, but reconstruction and attention maps come out black too"), down to the...

Hi @Luke2642 can I ask if you ever got this working?

Sorry, completely missed this, thanks @Luke2642!

Does this also apply when the model is shared between GPU and disk? My device map looks like ```python3 {'shared': 0, 'decoder.embed_tokens': 0, 'encoder.embed_tokens': 0, 'encoder.block.0': 0, 'encoder.block.1': 0, 'encoder.block.2':...

@ElleLeonne I'm not training though; what I mean is that during inference, model outputs are changing depending on how many outputs are being generated at a time, which shouldn't be...

Hi @lywinged I don't think the suggestions solve the underlying issue. If I follow both of them, different outputs are still being produced on `inputs = ["What country's flag is...

Thanks, though the suggested changes still don't fix the underlying problem I think. Like I mentioned here > Hi @lywinged I don't think the suggestions solve the underlying issue. If...

Sorry I'm not entirely sure what this code snippet > ``` > print(tokenizer.batch_decode(full_output, skip_special_tokens=False)[0]) > print('-' * 80) > print(tokenizer.batch_decode(full_output, skip_special_tokens=False)[1]) > ``` you shared is for. We're still getting...