Newest version with: onnx = Errors after I follow the instructions with pip install - HELP ? #99

New Issue

2021-08-23T04:30:48+02:00

AlonDan commented

2021-08-23 04:30:48 +02:00

(Migrated from github.com)

I've updated to latest SimSwap version and also pip install the latest commands based on the instruction suggested.

Now I get errors, I'm no programmer I don't know what it means or how to fix it.
It used to work before the all onnx update, now nothing works anymore...

Is there a way I can make onnx works? I would like to see the improvement of speed of course.

I'm using:

Windows 10
Anaconda
Cuda V10.2.89

This is the error I get: (it worked fine before the new onnx pip install)

UPDATE: (still need help)
I've used pip uninstalled onnxruntime-gpu (I understood 1.8 was installed by default and it's not for cuda 10.2)
and I tried to pip install onnxruntime-gpu==1.6 (it won't let me install 1.5 so I tried 1.5.1 and when it failed I uninstalled and reinstall again version 1.6) but I still get the same errors:

Can somebody please guide me step-by-step how to fix that and make onnx works?
Thanks ahead!

I've updated to latest SimSwap version and also pip install the latest commands based on the instruction suggested. Now I get errors, I'm no programmer I don't know what it means or how to fix it. It used to work before the all onnx update, now nothing works anymore... Is there a way I can make onnx works? I would like to see the improvement of speed of course. I'm using: - Windows 10 - Anaconda - Cuda V10.2.89 This is the error I get: (it worked fine before the new onnx pip install) ![2021-08-23_05-24-04](https://user-images.githubusercontent.com/21152334/130381805-f6b1f2c5-eb81-46cb-a75e-47f448db00b7.png) **UPDATE:** (still need help) I've used pip uninstalled onnxruntime-gpu (I understood 1.8 was installed by default and it's not for cuda 10.2) and I tried to pip install onnxruntime-gpu==1.6 (it won't let me install 1.5 so I tried 1.5.1 and when it failed I uninstalled and reinstall again version 1.6) but I still get the same errors: ![cmd_2021-08-23_05-43-28](https://user-images.githubusercontent.com/21152334/130382905-24513f68-a7e5-44c1-9ffc-4112c6a3ff85.png) Can somebody please guide me step-by-step how to fix that and make onnx works? Thanks ahead!

instant-high commented

2021-08-23 14:08:28 +02:00

(Migrated from github.com)

Uninstall onnxruntime-gpu and go back to install onnxruntime
Then it should work like before.

Uninstall onnxruntime-gpu and go back to install onnxruntime Then it should work like before.

bmc84 commented

2021-08-23 14:30:24 +02:00

(Migrated from github.com)

Hi,

I'm having issues as well. Following the same steps (including trying to uninstall onnxruntime-gpu).... now running the process just hangs. It gets as far as
.....

which_epoch: latest
-------------- End ----------------

then it just hangs. This is also happening in a brand new environment. I'm so perplexed at how this has broken it; perhaps you should remove this onnx GPU support if it's causing more grief than good.

[EDIT]: I notice under options, that I see: "onnx: None"

Does this somehow need to be enabled for this to work properly?

[EDIT EDIT]: It didn't hang: it took over an hour to load the model.... what's going on

Hi, I'm having issues as well. Following the same steps (including trying to uninstall onnxruntime-gpu).... now running the process just hangs. It gets as far as ..... ``` which_epoch: latest -------------- End ---------------- ``` then it just hangs. This is also happening in a brand new environment. I'm so perplexed at how this has broken it; perhaps you should remove this onnx GPU support if it's causing more grief than good. [EDIT]: I notice under options, that I see: "onnx: None" Does this somehow need to be enabled for this to work properly? [EDIT EDIT]: It didn't hang: it took *over an hour* to load the model.... what's going on

CongYuSanXun commented

2021-08-31 14:54:38 +02:00

(Migrated from github.com)

你好，

我也有问题。遵循相同的步骤（包括尝试卸载 onnxruntime-gpu）......现在运行该进程只是挂起。它达到
.....
which_epoch: latest
-------------- End ----------------
然后它就挂了。这也发生在一个全新的环境中。我对这如何破坏它感到非常困惑；也许你应该删除这个 onnx GPU 支持，如果它导致更多的悲伤而不是好处。

[编辑]：我注意到在选项下，我看到：“onnx：无”

这是否需要以某种方式启用才能正常工作？

[编辑编辑]：它没有挂起：加载模型花_了一个多小时_......发生了什么

--onnx ONNX ？

> 你好， > > 我也有问题。遵循相同的步骤（包括尝试卸载 onnxruntime-gpu）......现在运行该进程只是挂起。它达到 > ..... > > ``` > which_epoch: latest > -------------- End ---------------- > ``` > > 然后它就挂了。这也发生在一个全新的环境中。我对这如何破坏它感到非常困惑；也许你应该删除这个 onnx GPU 支持，如果它导致更多的悲伤而不是好处。 > > [编辑]：我注意到在选项下，我看到：“onnx：无” > > 这是否需要以某种方式启用才能正常工作？ > > [编辑编辑]：它没有挂起：加载模型花_了一个多小时_......发生了什么 --onnx ONNX ？

GreenHako commented

2021-09-02 07:29:40 +02:00

(Migrated from github.com)

i fixed this issue by installing cuda 10.1.243, then updating the simswap environment index in anaconda navigator.

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: CalvinBackup/neuralchen-SimSwap#99