SUDO-AI-3D / zero123plus

Code repository for Zero123++: a Single Image to Consistent Multi-view Diffusion Base Model.
Apache License 2.0
1.76k stars 123 forks source link

Why do the images I generated from different viewpoints look gray? #96

Closed Dumeowmeow closed 1 month ago

Dumeowmeow commented 2 months ago

like this:

1726303240487
eliphatfs commented 2 months ago

What code are you using?

Dumeowmeow commented 2 months ago

What code are you using?

The code I ran is as follows:

1726378650336
eliphatfs commented 2 months ago

Can you provide the code as a formatted code block in the comment so i may try to reproduce the behavior on my side?

Dumeowmeow commented 2 months ago

Can you provide the code as a formatted code block in the comment so i may try to reproduce the behavior on my side?

Thank you!Here is the code I used, actually, the only difference between this and img_to_mv.py is that the background is removed at the end.

import torch import requests from PIL import Image from diffusers import DiffusionPipeline, EulerAncestralDiscreteScheduler

pipeline = DiffusionPipeline.from_pretrained( "sudo-ai/zero123plus-v1.1", custom_pipeline="sudo-ai/zero123plus-pipeline", torch_dtype=torch.float16 )

pipeline.scheduler = EulerAncestralDiscreteScheduler.from_config( pipeline.scheduler.config, timestep_spacing='trailing' ) pipeline.to('cuda:0')

cond = Image.open(requests.get("https://d.skis.ltd/nrp/sample-data/lysol.png", stream=True).raw)

result = pipeline(cond, num_inference_steps=75).images[0]

import rembg result = rembg.remove(result) result.show() result.save("output.png")

eliphatfs commented 2 months ago

output I am running your code changing nothing and here is my result. Maybe there is a problem with the diffusers version?

Dumeowmeow commented 1 month ago

output I am running your code changing nothing and here is my result. Maybe there is a problem with the diffusers version?

Thank you! I reinstalled the environment and the problem is solved.