[HELP] Things went wrong after saving global profile with VS yuv420 chroma location left

[andrewschen]

Yes, now TensorRT / TensorRT RTX, FP16/32 all working just fine.

[Selur]

then it's likely a driver issue, since the only changes listed are upgrades,..

does it work if you use the newer version, open a terminal in the Hybrid/64bit/Vapoursynth folder and either call:

python -m pip install onnxconverter-common

(for non NVIDIA cards) or

python -m pip install nvidia-modelopt

(for NVIDIA cards).

Cu Selur

[andrewschen]

You mean using VapoursynthR77_torch_2026.06.16 + vs-mlrt_2026.06.06 right?

Results:

D:\APPD\Hybrid_dev_20260621-184241\64bit\Vapoursynth>python -m pip install nvidia-modelopt
Requirement already satisfied: nvidia-modelopt in .\Lib\site-packages (0.43.0)
Requirement already satisfied: ninja in .\Lib\site-packages (from nvidia-modelopt) (1.13.0)
Requirement already satisfied: numpy in .\Lib\site-packages (from nvidia-modelopt) (2.2.6)
Requirement already satisfied: nvidia-ml-py>=12 in .\Lib\site-packages (from nvidia-modelopt) (13.595.45)
Requirement already satisfied: packaging in .\Lib\site-packages (from nvidia-modelopt) (26.0)
Requirement already satisfied: setuptools>=80 in .\Lib\site-packages (from nvidia-modelopt) (82.0.1)
Requirement already satisfied: torch>=2.6 in .\Lib\site-packages (from nvidia-modelopt) (2.7.1+cu128)
Requirement already satisfied: tqdm in .\Lib\site-packages (from nvidia-modelopt) (4.67.1)
Requirement already satisfied: pulp in .\Lib\site-packages (from nvidia-modelopt) (3.3.0)
Requirement already satisfied: pydantic>=2.0 in .\Lib\site-packages (from nvidia-modelopt) (2.12.5)
Requirement already satisfied: regex in .\Lib\site-packages (from nvidia-modelopt) (2025.11.3)
Requirement already satisfied: rich in .\Lib\site-packages (from nvidia-modelopt) (14.2.0)
Requirement already satisfied: safetensors in .\Lib\site-packages (from nvidia-modelopt) (0.7.0)
Requirement already satisfied: scipy in .\Lib\site-packages (from nvidia-modelopt) (1.16.3)
Requirement already satisfied: PyYAML>=6.0 in .\Lib\site-packages (from nvidia-modelopt) (6.0.3)
Requirement already satisfied: omegaconf>=2.3.0 in .\Lib\site-packages (from nvidia-modelopt) (2.3.0)
Requirement already satisfied: antlr4-python3-runtime==4.9.* in .\Lib\site-packages (from omegaconf>=2.3.0->nvidia-modelopt) (4.9.3)
Requirement already satisfied: annotated-types>=0.6.0 in .\Lib\site-packages (from pydantic>=2.0->nvidia-modelopt) (0.7.0)
Requirement already satisfied: pydantic-core==2.41.5 in .\Lib\site-packages (from pydantic>=2.0->nvidia-modelopt) (2.41.5)
Requirement already satisfied: typing-extensions>=4.14.1 in .\Lib\site-packages (from pydantic>=2.0->nvidia-modelopt) (4.15.0)
Requirement already satisfied: typing-inspection>=0.4.2 in .\Lib\site-packages (from pydantic>=2.0->nvidia-modelopt) (0.4.2)
Requirement already satisfied: filelock in .\Lib\site-packages (from torch>=2.6->nvidia-modelopt) (3.20.0)
Requirement already satisfied: sympy>=1.13.3 in .\Lib\site-packages (from torch>=2.6->nvidia-modelopt) (1.14.0)
Requirement already satisfied: networkx in .\Lib\site-packages (from torch>=2.6->nvidia-modelopt) (3.6.1)
Requirement already satisfied: jinja2 in .\Lib\site-packages (from torch>=2.6->nvidia-modelopt) (3.1.6)
Requirement already satisfied: fsspec in .\Lib\site-packages (from torch>=2.6->nvidia-modelopt) (2025.12.0)
Requirement already satisfied: mpmath<1.4,>=1.1.0 in .\Lib\site-packages (from sympy>=1.13.3->torch>=2.6->nvidia-modelopt) (1.3.0)
Requirement already satisfied: MarkupSafe>=2.0 in .\Lib\site-packages (from jinja2->torch>=2.6->nvidia-modelopt) (2.1.5)
Requirement already satisfied: markdown-it-py>=2.2.0 in .\Lib\site-packages (from rich->nvidia-modelopt) (4.0.0)
Requirement already satisfied: pygments<3.0.0,>=2.13.0 in .\Lib\site-packages (from rich->nvidia-modelopt) (2.19.2)
Requirement already satisfied: mdurl~=0.1 in .\Lib\site-packages (from markdown-it-py>=2.2.0->rich->nvidia-modelopt) (0.1.2)
Requirement already satisfied: colorama in .\Lib\site-packages (from tqdm->nvidia-modelopt) (0.4.6)

[Selur]

torch- and vs-mlrt are separated, the above would only have changed anything when torch add-on wasn't used.
Seems like the latest vs-mlrt for some reason breaks on your setup.
=> you might want to report this to the vs-mlrt folks over at https://github.com/AmusementClub/vs-mlrt/

Cu Selur

[andrewschen]

I just thought about my laptop is Intel Arc Integrated / NV 5090 laptop, running in hybrid mode, so maybe provide the info of non-nvidia card will help?

D:\APPD\Hybrid_dev_20260621-184241\64bit\Vapoursynth>python -m pip install onnxconverter-common
Requirement already satisfied: onnxconverter-common in .\Lib\site-packages (1.16.0)
Requirement already satisfied: numpy in .\Lib\site-packages (from onnxconverter-common) (2.2.6)
Requirement already satisfied: onnx in .\Lib\site-packages (from onnxconverter-common) (1.19.1)
Requirement already satisfied: packaging in .\Lib\site-packages (from onnxconverter-common) (26.0)
Requirement already satisfied: protobuf>=3.20.2 in .\Lib\site-packages (from onnxconverter-common) (6.33.2)
Requirement already satisfied: typing_extensions>=4.7.1 in .\Lib\site-packages (from onnx->onnxconverter-common) (4.15.0)
Requirement already satisfied: ml_dtypes>=0.5.0 in .\Lib\site-packages (from onnx->onnxconverter-common) (0.5.4)

[andrewschen]

Got a response from the vs-mlrt author.

https://github.com/AmusementClub/vs-mlrt/issues/166

Guess that means need your new build of vs-mlrt?
I'll wait for the next build and test it out then.

[Selur]

Sounds like simply adding output_format=(0: fp32, 1: fp16) to the vs-mlrt Backend f:\Icalls in Hybrid itself should be enough,...
=> uploaded a new dev which should do this.

Cu Selur

[andrewschen]

Hybrid_dev_20260623-205841: not fixed

DPIRDenoise(mlrt) : TensorRT FP32 works ok, TensorRT FP16 and TensorRT RTX FP16/32 not.

SCUNet(mlrt) : TensorRT FP16/32 mode works ok, TensorRT RTX FP16/32 not.

DPIRDenoise(mlrt) TensorRT FP16 error log

2026-06-24 04:10:04.686
Plugin D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vstrt.dll is using API3 which is deprecated and will be removed shortly.
Plugin D:/APPD/Hybrid_dev_20260623-205841/64bit/vsfilters/SourceFilter/LSmashSource/LSMASHSource.dll is using API3 which is deprecated and will be removed shortly.
2026-06-24 04:10:05.009
Failed to evaluate the script:
Python exception: input format must be equal to output format

Traceback (most recent call last):
File "vapoursynth.pyx", line 3623, in vapoursynth._vpy_evaluate
File "vapoursynth.pyx", line 3624, in vapoursynth._vpy_evaluate
File "D:\APPD\HybridFiles\Temp\tempPreviewVapoursynthFile04_10_04_096.vpy", line 55, in
clip = vsmlrt.DPIR(clip, strength=5.000, model=1, backend=Backend.TRT(fp16=True,device_id=0,bf16=False,verbose=True,use_cuda_graph=False,num_streams=1,builder_optimization_level=1,engine_folder="D:/APPD/HybridFiles/Engine0606",output_format=0))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsmlrt.py", line 599, in DPIR
clip = inference_with_fallback(
^^^^^^^^^^^^^^^^^^^^^^^^
File "D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsmlrt.py", line 3110, in inference_with_fallback
raise e
File "D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsmlrt.py", line 3087, in inference_with_fallback
ret = _inference(
^^^^^^^^^^^
File "D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsmlrt.py", line 2904, in _inference
engine_path = trtexec(
^^^^^^^^
File "D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsmlrt.py", line 2087, in trtexec
convert_model(
File "D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsmlrt.py", line 1958, in convert_model
raise ValueError("input format must be equal to output format")
ValueError: input format must be equal to output format

DPIRDenoise(mlrt) TensorRT RTX FP16 error log

2026-06-24 04:10:39.487
Plugin D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vstrt_rtx.dll is using API3 which is deprecated and will be removed shortly.
Plugin D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vstrt.dll is using API3 which is deprecated and will be removed shortly.
Plugin D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsort.dll is using API3 which is deprecated and will be removed shortly.
Plugin D:/APPD/Hybrid_dev_20260623-205841/64bit/vsfilters/SourceFilter/LSmashSource/LSMASHSource.dll is using API3 which is deprecated and will be removed shortly.
2026-06-24 04:10:39.882
Failed to evaluate the script:
Python exception: Command '['D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt\\vsmlrt-cuda\\tensorrt_rtx', '--onnx=D:/APPD/HybridFiles/Engine0606\\drunet_color.onnx_3516927498_fp16.onnx', '--timingCacheFile=D:/APPD/HybridFiles/Engine0606\\d91c2312.engine.cache', '--device=0', '--saveEngine=D:/APPD/HybridFiles/Engine0606\\d91c2312.engine', '--useGpu', '--shapes=input:1x4x1080x1920', '--verbose', '--tacticSources=-CUDNN,+EDGE_MASK_CONVOLUTIONS', '--skipInference', '--builderOptimizationLevel=1']' returned non-zero exit status 1.

Traceback (most recent call last):
File "vapoursynth.pyx", line 3623, in vapoursynth._vpy_evaluate
File "vapoursynth.pyx", line 3624, in vapoursynth._vpy_evaluate
File "D:\APPD\HybridFiles\Temp\tempPreviewVapoursynthFile04_10_38_919.vpy", line 57, in
clip = vsmlrt.DPIR(clip, strength=5.000, model=1, backend=Backend.TRT_RTX(fp16=True,device_id=0,verbose=True,use_cuda_graph=False,num_streams=1,builder_optimization_level=1,engine_folder="D:/APPD/HybridFiles/Engine0606"))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsmlrt.py", line 599, in DPIR
clip = inference_with_fallback(
^^^^^^^^^^^^^^^^^^^^^^^^
File "D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsmlrt.py", line 3110, in inference_with_fallback
raise e
File "D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsmlrt.py", line 3087, in inference_with_fallback
ret = _inference(
^^^^^^^^^^^
File "D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsmlrt.py", line 3000, in _inference
engine_path = tensorrt_rtx(
^^^^^^^^^^^^^
File "D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsmlrt.py", line 2571, in tensorrt_rtx
subprocess.run(args, env=env, check=True, stdout=sys.stderr)
File "subprocess.py", line 571, in run
subprocess.CalledProcessError: Command '['D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt\\vsmlrt-cuda\\tensorrt_rtx', '--onnx=D:/APPD/HybridFiles/Engine0606\\drunet_color.onnx_3516927498_fp16.onnx', '--timingCacheFile=D:/APPD/HybridFiles/Engine0606\\d91c2312.engine.cache', '--device=0', '--saveEngine=D:/APPD/HybridFiles/Engine0606\\d91c2312.engine', '--useGpu', '--shapes=input:1x4x1080x1920', '--verbose', '--tacticSources=-CUDNN,+EDGE_MASK_CONVOLUTIONS', '--skipInference', '--builderOptimizationLevel=1']' returned non-zero exit status 1.

DPIRDenoise(mlrt) TensorRT RTX FP32 error log

2026-06-24 04:10:25.494
Plugin D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vstrt_rtx.dll is using API3 which is deprecated and will be removed shortly.
Plugin D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vstrt.dll is using API3 which is deprecated and will be removed shortly.
Plugin D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsort.dll is using API3 which is deprecated and will be removed shortly.
Plugin D:/APPD/Hybrid_dev_20260623-205841/64bit/vsfilters/SourceFilter/LSmashSource/LSMASHSource.dll is using API3 which is deprecated and will be removed shortly.
2026-06-24 04:10:25.880
Failed to evaluate the script:
Python exception: Command '['D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt\\vsmlrt-cuda\\tensorrt_rtx', '--onnx=D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt\\models\\dpir\\drunet_color.onnx', '--timingCacheFile=D:/APPD/HybridFiles/Engine0606\\5cac2f48.engine.cache', '--device=0', '--saveEngine=D:/APPD/HybridFiles/Engine0606\\5cac2f48.engine', '--useGpu', '--shapes=input:1x4x1080x1920', '--verbose', '--tacticSources=-CUDNN,+EDGE_MASK_CONVOLUTIONS', '--skipInference', '--builderOptimizationLevel=1']' returned non-zero exit status 1.

Traceback (most recent call last):
File "vapoursynth.pyx", line 3623, in vapoursynth._vpy_evaluate
File "vapoursynth.pyx", line 3624, in vapoursynth._vpy_evaluate
File "D:\APPD\HybridFiles\Temp\tempPreviewVapoursynthFile04_10_24_934.vpy", line 57, in
clip = vsmlrt.DPIR(clip, strength=5.000, model=1, backend=Backend.TRT_RTX(fp16=False,device_id=0,verbose=True,use_cuda_graph=False,num_streams=1,builder_optimization_level=1,engine_folder="D:/APPD/HybridFiles/Engine0606"))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsmlrt.py", line 599, in DPIR
clip = inference_with_fallback(
^^^^^^^^^^^^^^^^^^^^^^^^
File "D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsmlrt.py", line 3110, in inference_with_fallback
raise e
File "D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsmlrt.py", line 3087, in inference_with_fallback
ret = _inference(
^^^^^^^^^^^
File "D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsmlrt.py", line 3000, in _inference
engine_path = tensorrt_rtx(
^^^^^^^^^^^^^
File "D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsmlrt.py", line 2571, in tensorrt_rtx
subprocess.run(args, env=env, check=True, stdout=sys.stderr)
File "subprocess.py", line 571, in run
subprocess.CalledProcessError: Command '['D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt\\vsmlrt-cuda\\tensorrt_rtx', '--onnx=D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt\\models\\dpir\\drunet_color.onnx', '--timingCacheFile=D:/APPD/HybridFiles/Engine0606\\5cac2f48.engine.cache', '--device=0', '--saveEngine=D:/APPD/HybridFiles/Engine0606\\5cac2f48.engine', '--useGpu', '--shapes=input:1x4x1080x1920', '--verbose', '--tacticSources=-CUDNN,+EDGE_MASK_CONVOLUTIONS', '--skipInference', '--builderOptimizationLevel=1']' returned non-zero exit status 1.

SCUNet(mlrt) TensorRT RTX FP16 error log

2026-06-24 04:11:55.937
Plugin D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vstrt_rtx.dll is using API3 which is deprecated and will be removed shortly.
Plugin D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vstrt.dll is using API3 which is deprecated and will be removed shortly.
Plugin D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsort.dll is using API3 which is deprecated and will be removed shortly.
Plugin D:/APPD/Hybrid_dev_20260623-205841/64bit/vsfilters/SourceFilter/LSmashSource/LSMASHSource.dll is using API3 which is deprecated and will be removed shortly.
2026-06-24 04:11:56.289
Failed to evaluate the script:
Python exception: Command '['D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt\\vsmlrt-cuda\\tensorrt_rtx', '--onnx=D:/APPD/HybridFiles/Engine0606\\scunet_color_real_psnr.onnx_316040128_fp16_io.onnx', '--timingCacheFile=D:/APPD/HybridFiles/Engine0606\\827aa555.engine.cache', '--device=0', '--saveEngine=D:/APPD/HybridFiles/Engine0606\\827aa555.engine', '--useGpu', '--shapes=input:1x3x1080x1920', '--verbose', '--tacticSources=-CUDNN,+EDGE_MASK_CONVOLUTIONS', '--skipInference', '--builderOptimizationLevel=1']' returned non-zero exit status 1.

Traceback (most recent call last):
File "vapoursynth.pyx", line 3623, in vapoursynth._vpy_evaluate
File "vapoursynth.pyx", line 3624, in vapoursynth._vpy_evaluate
File "D:\APPD\HybridFiles\Temp\tempPreviewVapoursynthFile04_11_55_376.vpy", line 57, in
clip = vsmlrt.SCUNet(clip, model=3, backend=Backend.TRT_RTX(fp16=True,device_id=0,verbose=True,use_cuda_graph=False,num_streams=1,builder_optimization_level=1,engine_folder="D:/APPD/HybridFiles/Engine0606"))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsmlrt.py", line 1606, in SCUNet
clip = inference_with_fallback(
^^^^^^^^^^^^^^^^^^^^^^^^
File "D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsmlrt.py", line 3110, in inference_with_fallback
raise e
File "D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsmlrt.py", line 3087, in inference_with_fallback
ret = _inference(
^^^^^^^^^^^
File "D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsmlrt.py", line 3000, in _inference
engine_path = tensorrt_rtx(
^^^^^^^^^^^^^
File "D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsmlrt.py", line 2571, in tensorrt_rtx
subprocess.run(args, env=env, check=True, stdout=sys.stderr)
File "subprocess.py", line 571, in run
subprocess.CalledProcessError: Command '['D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt\\vsmlrt-cuda\\tensorrt_rtx', '--onnx=D:/APPD/HybridFiles/Engine0606\\scunet_color_real_psnr.onnx_316040128_fp16_io.onnx', '--timingCacheFile=D:/APPD/HybridFiles/Engine0606\\827aa555.engine.cache', '--device=0', '--saveEngine=D:/APPD/HybridFiles/Engine0606\\827aa555.engine', '--useGpu', '--shapes=input:1x3x1080x1920', '--verbose', '--tacticSources=-CUDNN,+EDGE_MASK_CONVOLUTIONS', '--skipInference', '--builderOptimizationLevel=1']' returned non-zero exit status 1.

SCUNet(mlrt) TensorRT RTX FP32 error log

2026-06-24 04:11:37.701
Plugin D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vstrt_rtx.dll is using API3 which is deprecated and will be removed shortly.
Plugin D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vstrt.dll is using API3 which is deprecated and will be removed shortly.
Plugin D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsort.dll is using API3 which is deprecated and will be removed shortly.
Plugin D:/APPD/Hybrid_dev_20260623-205841/64bit/vsfilters/SourceFilter/LSmashSource/LSMASHSource.dll is using API3 which is deprecated and will be removed shortly.
2026-06-24 04:11:38.052
Failed to evaluate the script:
Python exception: Command '['D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt\\vsmlrt-cuda\\tensorrt_rtx', '--onnx=D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt\\models\\scunet\\scunet_color_real_psnr.onnx', '--timingCacheFile=D:/APPD/HybridFiles/Engine0606\\368e98d3.engine.cache', '--device=0', '--saveEngine=D:/APPD/HybridFiles/Engine0606\\368e98d3.engine', '--useGpu', '--shapes=input:1x3x1080x1920', '--verbose', '--tacticSources=-CUDNN,+EDGE_MASK_CONVOLUTIONS', '--skipInference', '--builderOptimizationLevel=1']' returned non-zero exit status 1.

Traceback (most recent call last):
File "vapoursynth.pyx", line 3623, in vapoursynth._vpy_evaluate
File "vapoursynth.pyx", line 3624, in vapoursynth._vpy_evaluate
File "D:\APPD\HybridFiles\Temp\tempPreviewVapoursynthFile04_11_37_112.vpy", line 57, in
clip = vsmlrt.SCUNet(clip, model=3, backend=Backend.TRT_RTX(fp16=False,device_id=0,verbose=True,use_cuda_graph=False,num_streams=1,builder_optimization_level=1,engine_folder="D:/APPD/HybridFiles/Engine0606"))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsmlrt.py", line 1606, in SCUNet
clip = inference_with_fallback(
^^^^^^^^^^^^^^^^^^^^^^^^
File "D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsmlrt.py", line 3110, in inference_with_fallback
raise e
File "D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsmlrt.py", line 3087, in inference_with_fallback
ret = _inference(
^^^^^^^^^^^
File "D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsmlrt.py", line 3000, in _inference
engine_path = tensorrt_rtx(
^^^^^^^^^^^^^
File "D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt/vsmlrt.py", line 2571, in tensorrt_rtx
subprocess.run(args, env=env, check=True, stdout=sys.stderr)
File "subprocess.py", line 571, in run
subprocess.CalledProcessError: Command '['D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt\\vsmlrt-cuda\\tensorrt_rtx', '--onnx=D:/APPD/Hybrid_dev_20260623-205841/64bit/vs-mlrt\\models\\scunet\\scunet_color_real_psnr.onnx', '--timingCacheFile=D:/APPD/HybridFiles/Engine0606\\368e98d3.engine.cache', '--device=0', '--saveEngine=D:/APPD/HybridFiles/Engine0606\\368e98d3.engine', '--useGpu', '--shapes=input:1x3x1080x1920', '--verbose', '--tacticSources=-CUDNN,+EDGE_MASK_CONVOLUTIONS', '--skipInference', '--builderOptimizationLevel=1']' returned non-zero exit status 1.

[Selur]

Please, use code- not quote-blocks for code and outputs. Otherwise your post are unnecessarily hard to read.

clip = vsmlrt.DPIR(clip, strength=5.000, model=1, backend=Backend.TRT(fp16=True,device_id=0,bf16=False,verbose=True,use_cuda_graph=False,num_streams=1,builder_optimization_level=1,engine_folder="D:/APPD/HybridFiles/Engine0606",output_format=0))

This's a bug where I mixed up the output_format value.


[quote] clip = vsmlrt.SCUNet(clip, model=3, backend=Backend.TRT_RTX(fp16=True,device_id=0,verbose=True,use_cuda_graph=False,num_streams=1,builder_optimization_level=1,engine_folder="D:/APPD/HybridFiles/Engine0606"))[/quore]
that looks correct (no output_format for TRT_RTX).

=> Uploaded a dev which should fix the 'mlrt DPIR denoise output_format mix-up'. Maybe that helps with DPIRDenoise(mlrt) FP16.

What gpu are you using? (device_id=0)
What drivers are you using? (in case you are not using studio drivers, try those - I only use studio drivers)

Cu Selur

Ps.: Off to work maybe, back in 14hrs or so.

[andrewschen]

Got it, didn't aware of the code-block option, sorry about that.

Use default option of device_id=0, and checked that it's running on nv gpu with hw monitor app. Previously on vulkan develop driver 596.60, now on studio driver 610.62.

Hybrid_dev_2026.06.24-41554: now DPIRDenoise(mlrt) TensorRT FP16 works ok, rest is same as previous version.