search

iPERDance / iPERCore

Liquid Warping GAN with Attention: A Unified Framework for Human Image Synthesis
https://iperdance.github.io/work/impersonator-plus-plus.html
Apache License 2.0
2.42k stars 311 forks source link

RTX 3080 compatibility with pre-compiled/released windows version #98

Open RuwinX opened 3 years ago

RuwinX commented 3 years ago

Basically title? Tried following the installation instructions but I guess 3000 series cards aren't supported. The neural renderer portion of the script kept failing.

Tried installing the compiled release from here:

https://github.com/iPERDance/iPERCore/wiki/How-to-use-the-released-version-on-windows%3F

Ran the sample command and it still states that pytorch is incompatible with my GPU

torch\cuda__init__.py:125: UserWarning: GeForce RTX 3080 with CUDA capability sm_86 is not compatible with the current PyTorch installation. The current PyTorch install supports CUDA capabilities sm_37 sm_50 sm_60 sm_61 sm_70 sm_75 compute_37. If you want to use the GeForce RTX 3080 GPU with PyTorch, please check the instructions at https://pytorch.org/get-started/locally/

warnings.warn(incompatible_device_warn.format(device_name, capability, " ".join(arch_list), device_name))

tried going to https://pytorch.org/get-started/locally/ ran the commands it gave for both stable and nightly versions and then tried running the sample command. Same error shared above keeps popping up.

Commands run were 1) conda install pytorch torchvision torchaudio cudatoolkit=11.1 -c pytorch-nightly -c conda-forge 2) conda install pytorch torchvision torchaudio cudatoolkit=11.1 -c pytorch -c conda-forge

Is there some other configuration I have to add? I have CUDA 10.2 and CUDA 11.2 installed at the moment as well issue

LeoLee321 commented 3 years ago

Use Cuda11.0 or 11.1 or 11.2 and pytorch1.8

RuwinX commented 3 years ago

Use Cuda11.0 or 11.1 or 11.2 and pytorch1.8 Could elaborate on this a bit more? As I've already mentioned, I have tried installing the latest pytorch nightly build as well as the latest stable build [ using conda install pytorch torchvision torchaudio cudatoolkit=11.1 -c pytorch-nightly -c conda-forge ]

Even after installing, when I run it, it fails with the same error message. Have you managed to get it to work with an RTX 3000 series card?

LeoLee321 commented 3 years ago

Use Cuda11.0 or 11.1 or 11.2 and pytorch1.8 Could elaborate on this a bit more? As I've already mentioned, I have tried installing the latest pytorch nightly build as well as the latest stable build [ using conda install pytorch torchvision torchaudio cudatoolkit=11.1 -c pytorch-nightly -c conda-forge ]

Even after installing, when I run it, it fails with the same error message. Have you managed to get it to work with an RTX 3000 series card?

I can run it with RTX3070 image

malawadd commented 3 years ago

Use Cuda11.0 or 11.1 or 11.2 and pytorch1.8 Could elaborate on this a bit more? As I've already mentioned, I have tried installing the latest pytorch nightly build as well as the latest stable build [ using conda install pytorch torchvision torchaudio cudatoolkit=11.1 -c pytorch-nightly -c conda-forge ]

Even after installing, when I run it, it fails with the same error message. Have you managed to get it to work with an RTX 3000 series card?

I can run it with RTX3070 image

how did you set it up to use cuda 11.1 ?

aur42 commented 3 years ago

I also have this problem. Installed Cuda 11.0 + PyTorch 1.8.0 and get the same error message.