search

magic-research / magic-animate

[CVPR 2024] MagicAnimate: Temporally Consistent Human Image Animation using Diffusion Model
https://showlab.github.io/magicanimate/
BSD 3-Clause "New" or "Revised" License
10.36k stars 1.06k forks source link

Too much resource #29

Open bilalpinkrose opened 9 months ago

bilalpinkrose commented 9 months ago

Hi, thanks for this amazing project. I have issues about generating, when I use on RunPOD's cloud 80gb GPU I still can't get over 25 sampling steps. It gives this error on gradio: Error Unexpected token '<', " <!DOCTYPE "... is not valid JSON

this is terminal error: workspace/magic-animate/magicanimate/pipelines/pipeline_animation.py:624: FutureWarning: Accessing config attribute in_channels directly via 'UNet3DConditionModel' object attribute is deprecated. Please access 'in_channels' over 'UNet3DConditionModel's config object instead, e.g. 'unet.config.in_channels'. num_channels_latents = self.unet.in_channels

Priestru commented 9 months ago

Smth wrong. It consumes ~15GB at max for default resolution. Also those warnings do no harm and should not be related, I get plenty of similar ones.

zcxu-eric commented 9 months ago

Hi, thanks for this amazing project. I have issues about generating, when I use on RunPOD's cloud 80gb GPU I still can't get over 25 sampling steps. It gives this error on gradio: Error Unexpected token '<', " <!DOCTYPE "... is not valid JSON

this is terminal error: workspace/magic-animate/magicanimate/pipelines/pipeline_animation.py:624: FutureWarning: Accessing config attribute in_channels directly via 'UNet3DConditionModel' object attribute is deprecated. Please access 'in_channels' over 'UNet3DConditionModel's config object instead, e.g. 'unet.config.in_channels'. num_channels_latents = self.unet.in_channels

It shouldn't consume over 80G GPU memory, did you install and enable xformer correctly?

bilalpinkrose commented 9 months ago

Hi, thanks for this amazing project. I have issues about generating, when I use on RunPOD's cloud 80gb GPU I still can't get over 25 sampling steps. It gives this error on gradio: Error Unexpected token '<', " <!DOCTYPE "... is not valid JSON this is terminal error: workspace/magic-animate/magicanimate/pipelines/pipeline_animation.py:624: FutureWarning: Accessing config attribute in_channels directly via 'UNet3DConditionModel' object attribute is deprecated. Please access 'in_channels' over 'UNet3DConditionModel's config object instead, e.g. 'unet.config.in_channels'. num_channels_latents = self.unet.in_channels

It shouldn't consume over 80G GPU memory, did you install and enable xformer correctly?

Thanks for your reply, I installed requirements.txt correctly, what do you mean by enabling xformer?