01.01.2024, 18:44
Helluh and exactly that ↑ .. what topic say..
I get to see many ↓
These warning are generated using the resizer setting VSmlrt → backend TensorRT (and others aswell) using a pytorch converted to onxx model.
Also, (maybe important) have converted the same torch model twice. why? Well now, One is Fp32 → no errors runs fine wich is odd to me.
Because the other one ought to be faster, and it truly is because it set to fp16. But AlAS , it generates these messages unfortunately ..
noticed "lazy loading" which is another word for saying it conserves VRam (fp16) if iam not mistaken..
But, despite of those warnings everything runs fine ( the world moves on.. ) and hybride provide me with nice finished / polished output file ..
Fyi, errors start to show as soon i used the resizer VSmlrt .. did spot warnings with ESRgan too, but different kind.. Yet , jobs getting done and hybrid ejects nice output file ..
Is the something diz duck needs to be concerned about ehh ?
Yum Yum ↑ Food ↑ for thought i'd say
cheers,
TD
I get to see many ↓
apoursynth preview error: Registered plugin creator - ::BatchedNMSDynamic_TRT version 1 Registered plugin creator - ::BatchedNMS_TRT version 1 Registered plugin creator - ::BatchTilePlugin_TRT version 1 Registered plugin creator - ::Clip_TRT version 1 Registered plugin creator - ::CoordConvAC version 1 Registered plugin creator - ::CropAndResizeDynamic version 1 Registered plugin creator - ::CropAndResize version 1 Registered plugin creator - ::DecodeBbox3DPlugin version 1 Registered plugin creator - ::DetectionLayer_TRT version 1
Vapoursynth preview error: Registered plugin creator - ::EfficientNMS_Explicit_TF_TRT version 1 Registered plugin creator - ::EfficientNMS_Implicit_TF_TRT version 1 Registered plugin creator - ::EfficientNMS_ONNX_TRT version 1 Registered plugin creator - ::EfficientNMS_TRT version 1 Registered plugin creator - ::FlattenConcat_TRT version 1 Registered plugin creator - ::GenerateDetection_TRT version 1 Registered plugin creator - ::GridAnchor_TRT version 1 Registered plugin creator - ::GridAnchorRect_TRT version 1 Registered plugin creator - ::InstanceNormalization_TRT version 1 Registered plugin creator - ::InstanceNormalization_TRT version 2 Registered plugin creator - ::LReLU_TRT version 1 Registered plugin creator - ::ModulatedDeformConv2d version 1 Registered plugin creator - ::MultilevelCropAndResize_TRT version 1 Registered plugin creator - ::MultilevelProposeROI_TRT version 1 Registered plugin creator - ::MultiscaleDeformableAttnPlugin_TRT version 1 Registered plugin creator - ::NMSDynamic_TRT version 1 Registered plugin creator - ::NMS_TRT version 1 Registered plugin creator - ::Normalize_TRT version 1 Registered plugin creator - ::PillarScatterPlugin version 1 Registered plugin creator - ::PriorBox_TRT version 1 Registered plugin creator - ::ProposalDynamic version 1 Registered plugin creator - ::ProposalLayer_TRT version 1 Registered plugin creator - ::Proposal version 1 Registered plugin creator - ::PyramidROIAlign_TRT version 1 Registered plugin creator - ::Region_TRT version 1 Registered plugin creator - ::Reorg_TRT version 1 Registered plugin creator - ::ResizeNearest_TRT version 1 Registered plugin creator - ::ROIAlign_TRT version 1 Registered plugin creator - ::RPROI_TRT version 1 Registered plugin creator - ::ScatterND version 1 Registered plugin creator - ::SpecialSlice_TRT version 1 Registered plugin creator - ::Split version 1 Registered plugin creator - ::VoxelGeneratorPlugin version 1
Vapoursynth preview error: Loaded engine size: 34 MiB
Vapoursynth preview error: Deserialization required 16673 microseconds. [MemUsageChange] TensorRT-managed allocation in engine deserialization: CPU +1, GPU +32, now: CPU 1, GPU 32 (MiB)
Vapoursynth preview error: Total per-runner device persistent memory is 0 Total per-runner host persistent memory is 2072032
Vapoursynth preview error: Allocated activation device memory of size 477757440
Vapoursynth preview error: [MemUsageChange] TensorRT-managed allocation in IExecutionContext creation: CPU +2, GPU +455, now: CPU 3, GPU 487 (MiB) CUDA lazy loading is enabled.
Vapoursynth preview error: Plugin creator already registered - ::BatchedNMSDynamic_TRT version 1 Plugin creator already registered - ::BatchedNMS_TRT version 1 Plugin creator already registered - ::BatchTilePlugin_TRT version 1 Plugin creator already registered - ::Clip_TRT version 1 Plugin creator already registered - ::CoordConvAC version 1 Plugin creator already registered - ::CropAndResizeDynamic version 1 Plugin creator already registered - ::CropAndResize version 1 Plugin creator already registered - ::DecodeBbox3DPlugin version 1 Plugin creator already registered - ::DetectionLayer_TRT version 1 Plugin creator already registered - ::EfficientNMS_Explicit_TF_TRT version 1 Plugin creator already registered - ::EfficientNMS_Implicit_TF_TRT version 1 Plugin creator already registered - ::EfficientNMS_ONNX_TRT version 1 Plugin creator already registered - ::EfficientNMS_TRT version 1 Plugin creator already registered - ::FlattenConcat_TRT version 1 Plugin creator already registered - ::GenerateDetection_TRT version 1 Plugin creator already registered - ::GridAnchor_TRT version 1 Plugin creator already registered - ::GridAnchorRect_TRT version 1 Plugin creator already registered - ::InstanceNormalization_TRT version 1 Plugin creator already registered - ::InstanceNormalization_TRT version 2 Plugin creator already registered - ::LReLU_TRT version 1 Plugin creator already registered - ::ModulatedDeformConv2d version 1 Plugin creator already registered - ::MultilevelCropAndResize_TRT version 1 Plugin creator already registered - ::MultilevelProposeROI_TRT version 1 Plugin creator already registered - ::MultiscaleDeformableAttnPlugin_TRT version 1 Plugin creator already registered - ::NMSDynamic_TRT version 1 Plugin creator already registered - ::NMS_TRT version 1 Plugin creator already registered - ::Normalize_TRT version 1 Plugin creator already registered - ::PillarScatterPlugin version 1 Plugin creator already registered - ::PriorBox_TRT version 1 Plugin creator already registered - ::ProposalDynamic version 1 Plugin creator already registered - ::ProposalLayer_TRT version 1 Plugin creator already registered - ::Proposal version 1 Plugin creator already registered - ::PyramidROIAlign_TRT version 1 Plugin creator already registered - ::Region_TRT version 1 Plugin creator already registered - ::Reorg_TRT version 1 Plugin creator already registered - ::ResizeNearest_TRT version 1 Plugin creator already registered - ::ROIAlign_TRT version 1 Plugin creator already registered - ::RPROI_TRT version 1 Plugin creator already registered - ::ScatterND version 1 Plugin creator already registered - ::SpecialSlice_TRT version 1 Plugin creator already registered - ::Split version 1 Plugin creator already registered - ::VoxelGeneratorPlugin version 1
Vapoursynth preview error: Loaded engine size: 34 MiB
Vapoursynth preview error: Deserialization required 15769 microseconds. [MemUsageChange] TensorRT-managed allocation in engine deserialization: CPU +1, GPU +32, now: CPU 1, GPU 32 (MiB)
Vapoursynth preview error: Total per-runner device persistent memory is 0 Total per-runner host persistent memory is 2072032 Allocated activation device memory of size 477757440
Vapoursynth preview error: [MemUsageChange] TensorRT-managed allocation in IExecutionContext creation: CPU +2, GPU +455, now: CPU 3, GPU 487 (MiB) CUDA lazy loading is enabled.
These warning are generated using the resizer setting VSmlrt → backend TensorRT (and others aswell) using a pytorch converted to onxx model.
Also, (maybe important) have converted the same torch model twice. why? Well now, One is Fp32 → no errors runs fine wich is odd to me.
Because the other one ought to be faster, and it truly is because it set to fp16. But AlAS , it generates these messages unfortunately ..
noticed "lazy loading" which is another word for saying it conserves VRam (fp16) if iam not mistaken..
But, despite of those warnings everything runs fine ( the world moves on.. ) and hybride provide me with nice finished / polished output file ..
Fyi, errors start to show as soon i used the resizer VSmlrt .. did spot warnings with ESRgan too, but different kind.. Yet , jobs getting done and hybrid ejects nice output file ..
Is the something diz duck needs to be concerned about ehh ?
Yum Yum ↑ Food ↑ for thought i'd say
cheers,
TD