Not for me - so not a direct result of face detection.
Am I correct in my understanding that you have a major issue with Gigapixel 5.15 and despite issues being reported within hours of being made available you have not even posted a warning message on the download page never mind pull until the issue can be investigated?
I assume you have heard of the term âduty of careâ - if not I suggest you look it up - but please pull the update first.
Then Iâd suggest you investigate the common thread from a number of beta testers that they highlighted this issue but their reports disappeared.
And in case you are uncertain at this stage - yes I am a PAID licence holder of the software in question and yes it crashed for me also
Please upload your Logs here, add .txt as the extension, and go to Help-> Graphics info., press copy and paste in this thread along with the logs.
Please upload your logs here, add .txt as the extension, and go to Help-> Graphics info., press copy and paste in this thread along with the logs.
Could you explain the issue you are actually having along with going to Help-> Graphics info., press copy and paste in this thread.
As long as Iâm not using âFace Refinementâ, âAuto-detect Settingsâ on the right pane and âUse recommended optionâ in GAI settings window - I can use GAI 5.1.5.
1 Like
5.1.5 crash for me when I try to load an image.
AMD x2700
nvidia 2070 Super with latest drivers (not hotfix)
I canât do nothing in gigapixel.
If you manage to turn off âFace Refinementâ quickly (if itâs on), you should be able to use GAI next time.
I was searching where GAI saves its settings, and I only found this place:
I wonder if this helps you?
Thatâs it is it ?
Have a go at the customer because they didnât follow procedure?
I did mention the product and version I was having the issue with - like everyone else has had to work out themselves - uninstall and reinstall from the previous versions list.
And no Iâm not going to reinstall something that doesnât even start fully when launched - no Iâm not a beta tester - Iâm a customer who has paid good money for a product and now has something that isnât fit for purpose.
If as children you all want to sit around arguing over Symantecâs rather than doing something to resolve the issue - I was obviously wasting my time reporting back
And finally can you tell me how am I supposed to know from the update dialogue if the current update is simply the broken one or a replacement with a fix ?
Keith Armstrong
Please open GigaPixel and go to Help-> Graphics info, press Copy and paste the information here.
Why do you thing I am âhaving a go at youâ, I just asked you to explain the issue you are facing and paste the Graphics inf. from within the application here âŠ
What are you saying here, I didnât ask you to reinstall anything and I didnât mention Symantec. If you want help just do as i asked, it is a start to identify your issue. And I donât see any mention of your issue in this thread except you said it wasnât face detection.
I rolled back to 5.1.4, but here is the graphic info (from 5.1.4):
Application & Version: Topaz Gigapixel AI Version 5.1.4
Operating System: Windows 10 (10.0)
Graphics Hardware: GeForce RTX 2070 SUPER/PCIe/SSE2
OpenGL Driver: 3.3.0 NVIDIA 452.06
CPU RAM: 16310 MB
Video RAM: 8192 MB
Preview Limit: 8000 Pixels
AiDon this is not driver fault because Gigapixel 5.1.4 and older version works ok. When preview is active Gigapixel not responding and crash at 2% preview/calibration. When i disable preview, then crash when save image.
.
.
.
Application & Version: Topaz Gigapixel AI Version 5.1.5
Operating System: Windows 10 20H2 Beta 19042.487
Graphics Hardware: GeForce RTX 2080 Ti/PCIe/SSE2
OpenGL Driver: 3.3.0 NVIDIA 451.85/452.06/452.22/460.15
CPU RAM: 16320 MB
Video RAM: 11264 MB
Preview Limit: 8000 Pixels
Doesnât seem like you should have any issues, do you have logs? If logging is on they will be in the the log folder which you can find from the Help menu, add them all to a ZIP file, then upload to a file sharing site such as Google Drive, DropBox etc and post the link here.
You will find that calibration does take a comparatively long time and it will look like it is stuck on 2% especially when loading OpenVINO.
There is no guarantee that Beta drivers work as the NVIDIA component, TensorFlow, may not be updated as there may be changes.
If you do reload v5.1.5, before opening an image change your performance settings to NO/NO and run a enlargement to test.
https://drive.google.com/file/d/1_WGzLkClAyi5o5AYOI7bZb0_h9sWMuf3/view?usp=sharing
NVIDIA CUDA Deep Neural Network Library works correct on 5.1.4 but fails in 5.1.5
This is error from 5.1.5:
[CPU TF] [ 0 ] Start testing
2020-09-07T02:23:51.436 [Debug ] <TEUpsamplingWorker::cnnTF line #: 1737> Thread ID: 0x1e4c8a41b60 NN 4 X Tensorflow Operation Start
2020-09-07T02:23:51.436 [Debug ] <TCalNumBlocks::calNumBlocks line #: 30> Thread ID: 0x1e4c8a41b60 [input size] QSize(192, 192)
2020-09-07T02:23:51.436 [Debug ] <TCalNumBlocks::calNumBlocks line #: 31> Thread ID: 0x1e4c8a41b60 [numBlocks] QSize(1, 1)
2020-09-07T02:23:51.436 [Debug ] <TCalNumBlocks::calNumBlocks line #: 39> Thread ID: 0x1e4c8a41b60 [last block] boundary needs more pixels ( 0 , 0 )
2020-09-07T02:23:51.440 [Debug ] <TEUpsamplingWorker::cnnTF line #: 1770> Thread ID: 0x1e4c8a41b60 inTSMat size QSize(192, 192) inTSMat type âCV_16UC3â
2020-09-07T02:23:51.440 [Debug ] <TEUpsamplingWorker::cnnTF line #: 1771> Thread ID: 0x1e4c8a41b60 â New paddedInTSMat size QSize(192, 192) paddedInTSMat type âCV_16UC3â
2020-09-07T02:23:51.440 [Debug ] <TEUpsamplingWorker::cnnTF line #: 1782> Thread ID: 0x1e4c8a41b60 input block size 192 x 192
2020-09-07T02:23:52.952 [Warning] <TtfUtils::proc line #: 231> Thread ID: 0x1e4c8a41b60 ERROR!!! proc failed Unknown: 2 root error(s) found.
(0) Unknown: Failed to get convolution algorithm. This is probably because cuDNN failed to initialize, so try looking to see if a warning log message was printed above.
[[{{node tl_unet1x2x4x/Conv2D}}]]
[[netOutput4X/_7]]
(1) Unknown: Failed to get convolution algorithm. This is probably because cuDNN failed to initialize, so try looking to see if a warning log message was printed above.
[[{{node tl_unet1x2x4x/Conv2D}}]]
0 successful operations.
0 derived errors ignored.
2020-09-07T02:23:52.952 [Debug ] <TEUpsamplingWorker::cnnTF line #: 1784> Thread ID: 0x1e4c8a41b60 one block convert time: 1.512
2020-09-07T02:23:52.952 [Debug ] <TEUpsamplingWorker::cnnTF line #: 1785> Thread ID: 0x1e4c8a41b60 out block size 0 x 0
And This is form 5.1.4:
[CPU TF] [ 0 ] Start testing
2020-09-07T02:09:55.327 [Debug ] <TEUpsamplingWorker::cnnTF line #: 1724> Thread ID: 0x235c533c610 NN 4 X Tensorflow Operation Start
2020-09-07T02:09:55.327 [Debug ] <TCalNumBlocks::calNumBlocks line #: 30> Thread ID: 0x235c533c610 [input size] QSize(192, 192)
2020-09-07T02:09:55.327 [Debug ] <TCalNumBlocks::calNumBlocks line #: 31> Thread ID: 0x235c533c610 [numBlocks] QSize(1, 1)
2020-09-07T02:09:55.327 [Debug ] <TCalNumBlocks::calNumBlocks line #: 39> Thread ID: 0x235c533c610 [last block] boundary needs more pixels ( 0 , 0 )
2020-09-07T02:09:55.331 [Debug ] <TEUpsamplingWorker::cnnTF line #: 1757> Thread ID: 0x235c533c610 inTSMat size QSize(192, 192) inTSMat type âCV_16UC3â
2020-09-07T02:09:55.331 [Debug ] <TEUpsamplingWorker::cnnTF line #: 1758> Thread ID: 0x235c533c610 â New paddedInTSMat size QSize(192, 192) paddedInTSMat type âCV_16UC3â
2020-09-07T02:09:55.331 [Debug ] <TEUpsamplingWorker::cnnTF line #: 1769> Thread ID: 0x235c533c610 input block size 192 x 192
2020-09-07T02:09:57.023 [Debug ] <TEUpsamplingWorker::cnnTF line #: 1771> Thread ID: 0x235c533c610 one block convert time: 1.691
2020-09-07T02:09:57.023 [Debug ] <TEUpsamplingWorker::cnnTF line #: 1772> Thread ID: 0x235c533c610 out block size 768 x 768
2020-09-07T02:09:57.059 [Debug ] <TEUpsamplingWorker::cnnTF line #: 1812> Thread ID: 0x235c533c610 NN 4 X Operation End
2020-09-07T02:09:57.059 [Debug ] <TEUpsamplingWorker::cnnTF line #: 1813> Thread ID: 0x235c533c610 convert time: 1.732
2020-09-07T02:09:57.060 [Debug ] <TEUpsamplingWorker::doBenchmark line #: 920> Thread ID: 0x235c533c610
1 Like
Thank you, hopefully this will give more insight to Taylor as the processing options have been changed to add TensorFlow & CPU as a processing option.
And as I only have a GTX1050/4GB I donât see these types of issues that seem to be restricted to some of the RTX line which I see NVIDIA sending out hotfixes often for the gaming community.
1 Like
with the new update I canât see the blur settings on my screen. something changed about the spacing. I have a x280 ThinkPad, small screen, now I have to change my DPI way too small. please make the small adjustment back to what the screen spacing was.
for me, as long as i dont use âFace Refinementâ, âAuto-detect Settingsâ, and use âUse recommended optionâ in settings, it works for me, otherwise, it freezes at 2% preview
Please go to Help-> Graphics info., press Copy and paste the info here. Doesnât matter which version you are using.
