Tried the new Proteus Natural version but just get Error. Normal Proteus works fine on same video. This is Windows 10. I have an older graphics card (GTX 1080 Ti) but no errors related to that.
I have checked the logs and found 2 interesting error messages that mention C:\Users\alexa\topaz-labs\onnxruntime\onnxruntime. There is no user alexa on this PC so no idea where that came from. Can anyone make sense of these?
2026-02-01 14-46-06.301 Thread: 3000 Info FF Process Output: 10 e[1;31m2026-02-01 14:46:06.2783094 [E:onnxruntime:, sequential_executor.cc:572 onnxruntime::ExecuteKernel]
Non-zero status code returned while running DmlFusedNode_0_0 node. Name:‘DmlFusedNode_0_0’
Status Message: C:\Users\alexa\topaz-labs\onnxruntime\onnxruntime\core\providers\dml\DmlExecutionProvider\src\DmlGraphFusionHelper.cpp(1078)\onnxruntime.dll!00007FF952A2ED6C: (caller: 00007FF952ABEC2C)
Exception(2) tid(103c) 887A0006 The GPU will not respond to more commands, most likely because of an invalid command passed by the calling application.
2026-02-01 14:46:06 14372 CRITICAL: DML failed to prepare output: Non-zero status code returned while running DmlFusedNode_0_0 node. Name:‘DmlFusedNode_0_0’
Status Message: C:\Users\alexa\topaz-labs\onnxruntime\onnxruntime\core\providers\dml\DmlExecutionProvider\src\DmlGraphFusionHelper.cpp(1078)\onnxruntime.dll!00007FF952A2ED6C: (caller: 00007FF952ABEC2C)
Exception(2) tid(103c) 887A0006 The GPU will not respond to more commands, most likely because of an invalid command passed by the calling application.
Thanks for the tip on memory - I reduced the limit to 40% but get the same error
Proteus Natural has higher gpu memory requirements than previous versions of Proteus. When I reduced the “Max Memory %” preference significantly (40% or lower), TV downloaded a different model file with a smaller tile size that ran ok. This is on Windows.
I see you’ve updated your post with reference to your 1080 Ti. Gpu memory shouldn’t be an issue for you then (unlike my 1060 6GB).
I believe the status messages with folder paths for “Users/alexa” relate to the environment where this version of Topaz Video was compiled. I get simlar status messages in my logs with failed Proteus Natural enhancements at 100% memory. The successful enhancements at lower memory % don’t contain these messages.
I saw this error message already in older TVAI versions, I don’t think that has anything to do with the problem. My guess is that it’s the graphics card/driver.
Those messages are part of the logs for where errors in the code occur. They’re for error tracking, so of course they are meant to be there, and give the devs direction when issues occur.
So they do relate to a “problem” with the processing of Proteus Natural in the OPs case, we just don’t know what the cause is.
I presume I get those messages because of exceeding the vram capacity of my gpu.
yes i know, all i wanna say is, I get already this error “Users/alexa not found” in the log long time ago as Proteus Natural did not exist. There are several old posting around about the mystery user “Alexa”
