Blöde Frage, stellt man im Windows den HDR Mode auf ein oder aus ?
Beiträge von atmde
-
-
Okay danke Dir Harald, jetzt weiß ich wenigstens woran es liegt.
CUDA hängt fast durchgehend bei 100% wenn SVP auf 62 fps transkodiert.
RIFE AI war bei der Anschaffung auch kein Thema, dann ist es jetzt halt so.
-
Hab gerade probiert "Die Eiskönigin" und dann "Coco" zu transkodieren, da hat es direkt Error gebracht
Zitat20:33:01.207: ===== Starting mpv ======
20:33:01.207: Command line: C:\Program Files (x86)\SVP 4\mpv64\mpv.exe M:/Filme/Die Eiskönigin - Völlig unverfroren (2013)/Die Eiskönigin - Völlig unverfroren (2013) 1080p DTS.mkv --o=M:/Filme/Die Eiskönigin - Völlig unverfroren (2013)/Die Eiskönigin - Völlig unverfroren (2013) 1080p DTS.SVP.temporary.mkv --no-audio --no-sub --no-sub-auto --input-ipc-server=mpvencodepipe --input-media-keys=no --no-msg-color --vf=vapoursynth:[C:\Users\HTPC_User\AppData\Roaming\SVP4\scripts\ffff.py]:4:8 --of=matroska --ovc=hevc_nvenc --ovcopts=b=16332133,preset=slow,rc=vbr,maxrate=21231773,bufsize=32664266,time_base=1000/23976,threads=8
20:33:02.363: (+) Video --vid=1 (*) 'HD 1080p AVC' (h264 1920x856 23.976fps)
20:33:02.363: Audio --aid=1 --alang=ger (*) 'DTS 5.1 ' (dts 6ch 48000Hz)
20:33:02.363: Subs --sid=1 --slang=ger (*) 'complete' (dvd_subtitle)
20:33:03.427: vstrt: TensorRT version mismatch, built with 2135 but loaded with 2136; continue but fingers crossed...
20:33:03.487: [vapoursynth] Script evaluation failed:
20:33:03.487: [vapoursynth] Python exception: vsmlrt.RIFE: RIFE: multi must be at least 2
20:33:03.487: [vapoursynth]
20:33:03.487: [vapoursynth] Traceback (most recent call last):
20:33:03.487: [vapoursynth] File "src\cython\vapoursynth.pyx", line 2866, in vapoursynth._vpy_evaluate
20:33:03.487: [vapoursynth] File "src\cython\vapoursynth.pyx", line 2867, in vapoursynth._vpy_evaluate
20:33:03.487: [vapoursynth] File "C:\Users\HTPC_User\AppData\Roaming\SVP4\scripts\ffff.py", line 78, in <module>
20:33:03.487: [vapoursynth] smooth = interpolate(clip)
20:33:03.487: [vapoursynth] File "C:\Users\HTPC_User\AppData\Roaming\SVP4\scripts\ffff.py", line 57, in interpolate
20:33:03.487: [vapoursynth] smooth = RIFE(input_m,multi=rife_num,model=rife_mnum,backend=trt_backend)
20:33:03.487: [vapoursynth] File "C:\Program Files (x86)\SVP 4\rife\vsmlrt.py", line 981, in RIFE
20:33:03.487: [vapoursynth] raise ValueError(f'{func_name}: RIFE: multi must be at least 2')
20:33:03.487: [vapoursynth] ValueError: vsmlrt.RIFE: RIFE: multi must be at least 2
20:33:03.487: [vapoursynth]
20:33:03.542: (!!!) Intermediate file may be broken: M:\Filme\Die Eiskönigin - Völlig unverfroren (2013)\Die Eiskönigin - Völlig unverfroren (2013) 1080p DTS.SVP.temporary.mkv
20:33:03.542: ===== mpv exited with code 62097 =====
Ich habe aber mittlerweile kapiert, das ich auf der ersten Seite (Videoprofile) die gewünschte Bildfrequenz vorab einstellen muss, dann läuft auch das Transkodieren an.
Bisheriger Höchstwert um die 67.1 fps
Ansonsten läuft er so mit 66 fps konstant (bei 50-60% GPU Last).
-
So Titel ist durchgelaufen, seperater File wurde gespeichert und der soapige Effekt ist voll da.
Läuft mit 60fps.
Kann mir jemand was zu dem Log sagen ? Passt das so, was da drin steht ?
Muss ich hier was ändern, oder ist das normal ?
Kann man den "Con Air - Extended Edition (1997) 1080p AC3.SVP.temporary.mkv" file bedenkenlos löschen ?
Zitat21:49:35.338: [03/12/2023-21:49:35] [W] [TRT] CUDA lazy loading is not enabled. Enabli
21:49:35.354: ng it can significantly reduce device memory usage. See `CUDA_MODULE_LOADING` in https://docs.n
21:49:35.354: vidia.com/cuda/cuda-c-programming-guide/index.html#env-vars
Zitat21:49:31.624: ===== Starting mpv ======
21:49:31.624: Command line: C:\Program Files (x86)\SVP 4\mpv64\mpv.exe M:/Filme/Con Air - Extended Edition (1997)/Con Air - Extended Edition (1997) 1080p AC3.mkv --o=M:/Filme/Con Air - Extended Edition (1997)/Con Air - Extended Edition (1997) 1080p AC3.SVP.temporary.mkv --no-audio --no-sub --no-sub-auto --input-ipc-server=mpvencodepipe --input-media-keys=no --no-msg-color --vf=vapoursynth:[C:\Users\HTPC_User\AppData\Roaming\SVP4\scripts\ffff.py]:4:8 --of=matroska --ovc=hevc_nvenc --ovcopts=b=15645268,preset=slow,rc=vbr,maxrate=20338848,bufsize=31290536,time_base=1000/60000,threads=8
21:49:32.330: (+) Video --vid=1 (*) 'HD 1080p AVC' (h264 1920x820 25.000fps)
21:49:32.330: Audio --aid=1 --alang=ger (*) 'Dolby Digital 5.1 ' (ac3 6ch 48000Hz)
21:49:32.330: File tags:
21:49:32.330: Title: CONAIR · UNRATED · EXTENDED (1997)
21:49:32.455: vstrt: TensorRT version mismatch, built with 2135 but loaded with 2136; continue but fingers crossed...
21:49:32.517: C:/Program Files (x86)/SVP 4/rife\models\rife_v2\rife_v4.4_ensemble.onnx.1920x832_fp16_trt-8502_cudnn_I-fp16_O-fp16_NVIDIA-GeForce-RTX-3060-Ti_139fe355.engine not writable
21:49:32.517: change engine path to C:\Users\HTPC_U~1\AppData\Local\Temp\Program Files (x86)/SVP 4/rife\models\rife_v2\rife_v4.4_ensemble.onnx.1920x832_fp16_trt-8502_cudnn_I-fp16_O-fp16_NVIDIA-GeForce-RTX-3060-Ti_139fe355.engine
21:49:32.614: &&&& RUNNING TensorRT.trtexec [TensorRT v8501] # C:/Program Files (x86)/SVP 4/rife\vsmlrt-cuda\trtexec --onnx=C:/Program Files (x86)/SVP 4/rife\models\rife_v2\rife_v4.4_ensemble.onnx --timingCacheFile=C:\Users\HTPC_U~1\AppData\Local\Temp\Program Files (x86)/SVP 4/rife\models\rife_v2\rife_v4.4_ensemble.onnx.1920x832_fp16_trt-8502_cudnn_I-fp16_O-fp16_NVIDIA-GeForce-RTX-3060-Ti_139fe355.engine.cache --device=0 --saveEngine=C:\Users\HTPC_U~1\AppData\Local\Temp\Program Files (x86)/SVP 4/rife\models\rife_v2\rife_v4.4_ensemble.onnx.1920x832_fp16_trt-8502_cudnn_I-fp16_O-fp16_NVIDIA-GeForce-RTX-3060-Ti_139fe355.engine --shapes=input:1x7x832x1920 --fp16 --tacticSources=-CUBLAS,-CUBLAS_LT --useCudaGraph --noDataTransfers --inputIOFormats=fp16:chw --outputIOFormats=fp16:chw
21:49:32.614: [03/12/2023-21:49:32] [I] === Model Options ===
21:49:32.614: [03/12/2023-21:49:32] [I] Format: ONNX
21:49:32.614: [03/12/2023-21:49:32] [I] Model: C:/Program Files (x86)/SVP 4/rife\models\rife_v2\rife_v4.4_ensemble.onnx
21:49:32.614: [03/12/2023-21:49:32] [I] Output:
21:49:32.614: [03/12/2023-21:49:32] [I] === Build Options ===
21:49:32.614: [03/12/2023-21:49:32] [I] Max batch: explicit batch
21:49:32.614: [03/12/2023-21:49:32] [I] Memory Pools: workspace: default, dlaSRAM: default, dlaLocalDRAM: default, dlaGlobalDRAM: default
21:49:32.614: [03/12/2023-21:49:32] [I] minTiming: 1
21:49:32.614: [03/12/2023-21:49:32] [I] avgTiming: 8
21:49:32.614: [03/12/2023-21:49:32] [I] Precision: FP32+FP16
21:49:32.614: [03/12/2023-21:49:32] [I] LayerPrecisions:
21:49:32.614: [03/12/2023-21:49:32] [I] Calibration:
21:49:32.614: [03/12/2023-21:49:32] [I] Refit: Disabled
21:49:32.614: [03/12/2023-21:49:32] [I] Sparsity: Disabled
21:49:32.614: [03/12/2023-21:49:32] [I] Safe mode: Disabled
21:49:32.614: [03/12/2023-21:49:32] [I] DirectIO mode: Disabled
21:49:32.614: [03/12/2023-21:49:32] [I] Restricted mode: Disabled
21:49:32.614: [03/12/2023-21:49:32] [I] Build only: Disabled
21:49:32.614: [03/12/2023-21:49:32] [I] Save engine: C:\Users\HTPC_U~1\AppData\Local\Temp\Program Files (x86)/SVP 4/rife\models\rife_v2\rife_v4.4_ensemble.onnx.1920x832_fp16_trt-8502_cudnn_I-fp16_O-fp16_NVIDIA-GeForce-RTX-3060-Ti_139fe355.engine
21:49:32.614: [03/12/2023-21:49:32] [I] Load engine:
21:49:32.614: [03/12/2023-21:49:32] [I] Profiling verbosity: 0
21:49:32.614: [03/12/2023-21:49:32] [I] Tactic sources: cublas [OFF], cublasLt [OFF],
21:49:32.614: [03/12/2023-21:49:32] [I] timingCacheMode: global
21:49:32.614: [03/12/2023-21:49:32] [I] timingCacheFile: C:\Users\HTPC_U~1\AppData\Local\Temp\Program Files (x86)/SVP 4/rife\models\rife_v2\rife_v4.4_ensemble.onnx.1920x832_fp16_trt-8502_cudnn_I-fp16_O-fp16_NVIDIA-GeForce-RTX-3060-Ti_139fe355.engine.cache
21:49:32.614: [03/12/2023-21:49:32] [I] Heuristic: Disabled
21:49:32.614: [03/12/2023-21:49:32] [I] Preview Features: Use default preview flags.
21:49:32.614: [03/12/2023-21:49:32] [I] Input(s): fp16:chw
21:49:32.614: [03/12/2023-21:49:32] [I] Output(s): fp16:chw
21:49:32.614: [03/12/2023-21:49:32] [I] Input build shape: input=1x7x832x1920+1x7x832x1920+1x7x832x1920
21:49:32.614: [03/12/2023-21:49:32] [I] Input calibration shapes: model
21:49:32.614: [03/12/2023-21:49:32] [I] === System Options ===
21:49:32.614: [03/12/2023-21:49:32] [I] Device: 0
21:49:32.614: [03/12/2023-21:49:32] [I] DLACore:
21:49:32.614: [03/12/2023-21:49:32] [I] Plugins:
21:49:32.614: [03/12/2023-21:49:32] [I] === Inference Options ===
21:49:32.614: [03/12/2023-21:49:32] [I] Batch: Explicit
21:49:32.614: [03/12/2023-21:49:32] [I] Input inference shape: input=1x7x832x1920
21:49:32.614: [03/12/2023-21:49:32] [I] Iterations: 10
21:49:32.614: [03/12/2023-21:49:32] [I] Duration: 3s (+ 200ms warm up)
21:49:32.614: [03/12/2023-21:49:32] [I] Sleep time: 0ms
21:49:32.614: [03/12/2023-21:49:32] [I] Idle time: 0ms
21:49:32.614: [03/12/2023-21:49:32] [I] Streams: 1
21:49:32.614: [03/12/2023-21:49:32] [I] ExposeDMA: Disabled
21:49:32.614: [03/12/2023-21:49:32] [I] Data transfers: Disabled
21:49:32.614: [03/12/2023-21:49:32] [I] Spin-wait: Disabled
21:49:32.614: [03/12/2023-21:49:32] [I] Multithreading: Disabled
21:49:32.614: [03/12/2023-21:49:32] [I] CUDA Graph: Enabled
21:49:32.614: [03/12/2023-21:49:32] [I] Separate profiling: Disabled
21:49:32.614: [03/12/2023-21:49:32] [I] Time Deserialize: Disabled
21:49:32.614: [03/12/2023-21:49:32] [I] Time Refit: Disabled
21:49:32.614: [03/12/2023-21:49:32] [I] NVTX verbosity: 0
21:49:32.614: [03/12/2023-21:49:32] [I] Persistent Cache Ratio: 0
21:49:32.614: [03/12/2023-21:49:32] [I] Inputs:
21:49:32.614: [03/12/2023-21:49:32] [I] === Reporting Options ===
21:49:32.614: [03/12/2023-21:49:32] [I] Verbose: Disabled
21:49:32.614: [03/12/2023-21:49:32] [I] Averages: 10 inferences
21:49:32.614: [03/12/2023-21:49:32] [I] Percentiles: 90,95,99
21:49:32.614: [03/12/2023-21:49:32] [I] Dump refittable layers:Disabled
21:49:32.614: [03/12/2023-21:49:32] [I] Dump output: Disabled
21:49:32.614: [03/12/2023-21:49:32] [I] Profile: Disabled
21:49:32.614: [03/12/2023-21:49:32] [I] Export timing to JSON file:
21:49:32.614: [03/12/2023-21:49:32] [I] Export output to JSON file:
21:49:32.614: [03/12/2023-21:49:32] [I] Export profile to JSON file:
21:49:32.614: [03/12/2023-21:49:32] [I]
21:49:32.629: [03/12/2023-21:49:32] [I] === Device Information ===
21:49:32.629: [03/12/2023-21:49:32] [I] Selected Device: NVIDIA GeForce RTX 3060 Ti
21:49:32.629: [03/12/2023-21:49:32] [I] Compute Capability: 8.6
21:49:32.629: [03/12/2023-21:49:32] [I] SMs: 38
21:49:32.629: [03/12/2023-21:49:32] [I] Compute Clock Rate: 1.665 GHz
21:49:32.629: [03/12/2023-21:49:32] [I] Device Global Memory: 8191 MiB
21:49:32.629: [03/12/2023-21:49:32] [I] Shared Memory per SM: 100 KiB
21:49:32.629: [03/12/2023-21:49:32] [I] Memory Bus Width: 256 bits (ECC disabled)
21:49:32.629: [03/12/2023-21:49:32] [I] Memory Clock Rate: 7.001 GHz
21:49:32.629: [03/12/2023-21:49:32] [I]
21:49:32.629: [03/12/2023-21:49:32] [I] TensorRT version: 8.5.1
21:49:32.973: [03/12/2023-21:49:32] [I] [TRT] [MemUsageChange] Init CUDA: CPU +446, GPU +0, now: CPU 10935, GPU 1183 (MiB)
21:49:35.323: [03/12/2023-21:49:35] [I] [TRT] [MemUsageChange] Init builder kernel library: CPU +483, GPU +118, now: CPU 11787, GPU 1301 (MiB)
21:49:35.338: [03/12/2023-21:49:35] [W] [TRT] CUDA lazy loading is not enabled. Enabli
21:49:35.354: ng it can significantly reduce device memory usage. See `CUDA_MODULE_LOADING` in https://docs.n
21:49:35.354: vidia.com/cuda/cuda-c-programming-guide/index.html#env-vars
21:49:35.354: [03/12/2023-21:49:35] [I] Start parsing network model
21:49:35.369: [03/12/2023-21:49:35] [I] [TRT] ----------------------------------------------------------------
21:49:35.369: [03/12/2023-21:49:35] [I] [TRT] Input filename: C:/Program Files (x86)/SVP 4/rife\models\rife_v2\rife_v4.4_ensemble.onnx
21:49:35.369: [03/12/2023-21:49:35] [I] [TRT] ONNX IR version: 0.0.8
21:49:35.369: [03/12/2023-21:49:35] [I] [TRT] Opset version: 16
21:49:35.369: [03/12/2023-21:49:35] [I] [TRT] Producer name: pytorch
21:49:35.369: [03/12/2023-21:49:35] [I] [TRT] Producer version: 2.0.0
21:49:35.369: [03/12/2023-21:49:35] [I] [TRT] Domain:
21:49:35.369: [03/12/2023-21:49:35] [I] [TRT] Model version: 0
21:49:35.369: [03/12/2023-21:49:35] [I] [TRT] Doc string:
21:49:35.369: [03/12/2023-21:49:35] [I] [TRT] ----------------------------------------------------------------
21:49:35.385: [03/12/2023-21:49:35] [W] [TRT] onnx2trt_utils.cpp:377: Your ONNX model has been generated with INT64 weights, while TensorRT does not natively support INT64. Attempting to cast down to INT32.
21:49:35.385: [03/12/2023-21:49:35] [W] [TRT] onnx2trt_utils.cpp:403: One or more weights outside the range of INT32 was clamped
21:49:35.385: [03/12/2023-21:49:35] [W] [TRT] onnx2trt_utils.cpp:403: One or more weights outside the range of INT32 was clamped
21:49:35.401: [03/12/2023-21:49:35] [W] [TRT] onnx2trt_utils.cpp:403: One or more weights outside the range of INT32 was clamped
21:49:35.401: [03/12/2023-
21:49:35.401: 21:49:35] [W] [TRT] onnx2trt_utils.cpp:403: One or more weights outside the range of INT32 was clamped
21:49:35.916: [03/12/2023-21:49:35] [W] [TRT] onnx2trt_utils.cpp:403: One or more weights outside the range of INT32 was clamped
21:49:35.932: [03/12/2023-21:49:35] [I] Finish parsing network model
21:49:35.932: [03/12/2023-21:49:35] [W] Could not read timing cache from: C:\Users\HTPC_U~1\AppData\Local\Temp\Program Files (x86)/SVP 4/rife\models\rife_v2\rife_v4.4_ensemble.onnx.1920x832_fp16_trt-8502_cudnn_I-fp16_O-fp16_NVIDIA-GeForce-RTX-3060-Ti_139fe355.engine.cache. A new timing cache will be generated and written.
21:49:37.087: [03/12/2023-21:49:37] [I] [TRT] [MemUsageChange] Init cuDNN: CPU +1091, GPU +406, now: CPU 12525, GPU 1707 (MiB)
21:49:37.087: [03/12/2023-21:49:37] [I] [TRT] Global timing cache in use. Profiling results in this builder pass will be stored.
21:52:48.071: [03/12/2023-21:52:48] [I] [TRT] Total Activation Memory: 9759461888
21:52:48.071: [03/12/2023-21:52:48] [I] [TRT] Detected 4 inputs and 1 output network tensors.
21:52:48.258: [03/12/2023-21:52:48] [I] [TRT] Total Host Persistent Memory: 202848
21:52:48.258: [03/12/2023-21:52:48] [I] [TRT] Total Device Persistent Memory: 16285696
21:52:48.258: [03/12/2023-21:52:48] [I] [TRT] Total Scratch Memory: 19169280
21:52:48.258: [03/12/2023-21:52:48] [I] [TRT] [MemUsageStats] Peak memory usage of TRT CPU/GPU memory allocators: CPU 19 MiB, GPU 4715 MiB
21:52:48.258: [03/12/2023-21:52:48] [I] [TRT] [BlockAssignment] Started assigning block shifts. This will take 154 steps to complete.
21:52:48.274: [03/12/2023-21:52:48] [I] [TRT] [BlockAssignment] Algorithm ShiftNTopDown took 12.6033ms to assign 13 blocks to 154 nodes requiring 233227264 bytes.
21:52:48.274: [03/12/2023-21:52:48] [I] [TRT] Total Activation Memory: 233227264
21:52:48.383: [03/12/2023-21:52:48] [W] [TRT] TensorRT encountered issues when converting weights between types and that could affect accuracy.
21:52:48.383: [03/12/2023-21:52:48] [W] [TRT] If this is not the desired behavior, please modify the weights or retrain with regularization to adjust the magnitude of the weights.
21:52:48.383: [03/12/2023-21:52:48] [W] [TRT] Check verbose logs for the list of affected weights.
21:52:48.383: [03/12/2023-21:
21:52:48.383: 52:48] [W] [TRT] - 94 weights are affected by this issue: Detected subnormal FP16 values.
21:52:48.383: [03/12/2023-21:52:48] [W] [TRT] - 10 weights are affected by this issue: Detected values less than smallest positive FP16 subnormal value and converted them to the FP16 minimum subnormalized value.
21:52:48.383: [03/12/2023-21:52:48] [I] [TRT] [MemUsageChange] TensorRT-managed allocation in building engine: CPU +19, GPU +25, now: CPU 19, GPU 25 (MiB)
21:52:48.399: [03/12/2023-21:52:48] [I] [TRT] [MemUsageChange] Init CUDA: CPU +0, GPU +0, now: CPU 13471, GPU 1933 (MiB)
21:52:48.399: [03/12/2023-21:52:48] [W] [TRT] CUDA lazy loading is not enabled. Enabling it can significantly reduce device memory usage. See `CUDA_MODULE_LOADING` in https://docs.nvidia.com/cuda/c…guide/index.html#env-vars
21:52:48.399: [03/12/2023-21:52:48] [I] Saved 190208 bytes of timing cache to C:\Users\HTPC_U~1\AppData\Local\Temp\Program Files (x86)/SVP 4/rife\models\rife_v2\rife_v4.4_ensemble.onnx.1920x832_fp16_trt-8502_cudnn_I-fp16_O-fp16_NVIDIA-GeForce-RTX-3060-Ti_139fe355.engine.cache
21:52:48.399: [03/12/2023-21:52:48] [I] Engine built in 195.767 sec.
21:52:48.555: [03/12/2023-21:52:48] [I] [TRT] Loaded engine size: 10 MiB
21:52:48.571: [03/12/2023-21:52:48] [I] [TRT] [MemUsageChange] TensorRT-managed allocation in engine deserialization: CPU +0, GPU +25, now: CPU 0, GPU 25 (MiB)
21:52:48.571: [03/12/2023-21:52:48] [I] Engine deserialized in 0.0090543 sec.
21:52:48.586: [03/12/2023-21:52:48] [I] [TRT] [MemUsageChange] TensorRT-managed allocation in IExecutionContext creation: CPU +0, GPU +238, now: CPU 0, GPU 263 (MiB)
21:52:48.586: [03/12/2023-21:52:48] [W] [TRT] CUDA lazy loading is not enabled. Enabling it can significantly reduce device memory usage. See `CUDA_MODULE_LOADING` in https://docs.nvidia.com/cuda/c…guide/index.html#env-vars
21:52:48.586: [03/12/2023-21:52:48] [I] Setting persistentCacheLimit to 0 bytes.
21:52:48.586: [03/12/2023-21:52:48] [I] Using random values for input input
21:52:48.727: [03/12/2023-21:52:48] [I] Created input binding for input with dimensions 1x7x832x1920
21:52:48.727: [03/12/2023-21:52:48] [I] Using random values for output output
21:52:48.727: [03/12/2023-21:52:48] [I] Created output binding for output with dimensions 1x3x832x1920
21:52:48.727: [03/12/2023-21:52:48] [I] Starting inference
21:52:51.983: [03/12/2023-21:52:51] [I] Warmup completed 12 queries over 200 ms
21:52:51.983: [03/12/2023-21:52:51] [I] Timing trace has 214 queries over 3.03613 s
21:52:51.983: [03/12/2023-21:52:51] [I]
21:52:51.983: [03/12/2023-21:52:51] [I] === Trace details ===
21:52:51.983: [03/12/2023-21:52:51] [I] Trace averages of 10 runs:
21:52:51.983: [03/12/2023-21:52:51] [I] Average on 10 runs - GPU latency: 14.2143 ms - Host latency: 14.2143 ms (enqueue 0.0871597 ms)
21:52:51.983: [03/12/2023-21:52:51] [I] Average on 10 runs - GPU latency: 14.0171 ms - Host latency: 14.0171 ms (enqueue 0.0539063 ms)
21:52:51.983: [03/12/2023-21:52:51] [I] Average on 10 runs - GPU latency: 14.1252 ms - Host latency: 14.1252 ms (enqueue 0.069458 ms)
21:52:51.983: [03/12/2023-21:52:51] [I] Average on 10 runs - GPU latency: 14.1378 ms - Host latency: 14.1378 ms (enqueue 0.0795288 ms)
21:52:51.983: [03/12/2023-21:52:51] [I] Average on 10 runs - GPU latency: 14.152 ms - Host latency: 14.152 ms (enqueue 0.0790649 ms)
21:52:51.983: [03/12/2023-21:52:51] [I] Average on 10 runs - GPU latency: 14.1415 ms - Host latency: 14.1415 ms (enqueue 0.0941895 ms)
21:52:51.983: [03/12/2023-21:52:51] [I] Average on 10 runs - GPU latency: 14.2885 ms - Host latency: 14.2885 ms (enqueue 0.0805664 ms)
21:52:51.983: [03/12/2023-21:52:51] [I] Average on 10 runs - GPU latency: 14.2182 ms - Host latency: 14.2182 ms (enqueue 0.0821167 ms)
21:52:51.983: [03/12/2023-21:52:51] [I] Average on 10 runs - GPU latency: 14.0829 ms - Host latency: 14.0829 ms (enqueue 0.0558594 ms)
21:52:51.983: [03/12/2023-21:52:51] [I] Average on 10 runs - GPU latency: 14.023 ms - Host latency: 14.023 ms (enqueue 0.0457031 ms)
21:52:51.983: [03/12/2023-21:52:51] [I] Average on 10 runs - GPU latency: 14.2037 ms - Host latency: 14.2037 ms (enqueue 0.0845825 ms)
21:52:51.983: [03/12/2023-21:52:51] [I] Average on 10 runs - GPU latency: 14.1831 ms - Host latency: 14.1831 ms (enqueue 0.0817749 ms)
21:52:51.983: [03/12/2023-21:52:51] [I] Average on 10 runs - GPU latency: 14.2008 ms - Host latency: 14.2008 ms (enqueue 0.0818237 ms)
21:52:51.983: [03/12/2023-21:52:51] [I] Average on 10 runs - GPU latency: 14.1933 ms - Host latency: 14.1933 ms (enqueue 0.0807739 ms)
21:52:51.983: [03/12/2023-21:52:51] [I] Average on 10 runs - GPU latency: 14.2031 ms - Host latency: 14.2031 ms (enqueue 0.0953369 ms)
21:52:51.983: [03/12/2023-21:52:51] [I] Average on 10 runs - GPU latency: 14.1423 ms - Host latency: 14.1423 ms (enqueue 0.0489014 ms)
21:52:51.983: [03/12/2023-21:52:51] [I] Average on 10 runs - GPU latency: 14.2208 ms - Host latency: 14.2208 ms (enqueue 0.0665527 ms)
21:52:51.983: [03/12/2023-21:52:51] [I] Average on 10 runs - GPU latency: 14.1942 ms - Host latency: 14.1942 ms (enqueue 0.0705078 ms)
21:52:51.983: [03/12/2023-21:52:51] [I] Average on 10 runs - GPU latency: 14.214 ms - Host latency: 14.214 ms (enqueue 0.0716553 ms)
21:52:51.983: [03/12/2023-21:52:51] [I] Average on 10 runs - GPU latency: 14.2147 ms - Host latency: 14.2147 ms (enqueue 0.0909668 ms)
21:52:51.983: [03/12/2023-21:52:51] [I] Average on 10 runs - GPU latency: 14.2426 ms - Host latency: 14.2426 ms (enqueue 0.0773682 ms)
21:52:51.983: [03/12/2023-21:52:51] [I]
21:52:51.983: [03/12/2023-21:52:51] [I] === Performance summary ===
21:52:51.983: [03/12/2023-21:52:51] [I] Throughput: 70.4844 qps
21:52:51.983: [03/12/2023-21:52:51] [I] Latency: min = 13.8905 ms, max = 15.1262 ms, mean = 14.1728 ms, median = 14.1734 ms, percentile(90%) = 14.2683 ms, percentile(95%) = 14.302 ms, percentile(99%) = 14.4467 ms
21:52:51.983: [03/12/2023-21:52:51] [I] Enqueue Time: min = 0.0153809 ms, max = 0.233643 ms, mean = 0.0750602 ms, median = 0.0795898 ms, percentile(90%) = 0.0843506 ms, percentile(95%) = 0.0913086 ms, percentile(99%) = 0.193237 ms
21:52:51.983: [03/12/2023-21:52:51] [I] H2D Latency: min = 0 ms, max = 0 ms, mean = 0 ms, median = 0 ms, percentile(90%) = 0 ms, percentile(95%) = 0 ms, percentile(99%) = 0 ms
21:52:51.983: [03/12/2023-21:52:51] [I] GPU Compute Time: min = 13.8905 ms, max = 15.1262 ms, mean = 14.1728 ms, median = 14.1734 ms, percentile(90%) = 14.2683 ms, percentile(95%) = 14.302 ms, percentile(99%) = 14.4467 ms
21:52:51.983: [03/12/2023-21:52:51] [I] D2H Latency: min = 0 ms, max = 0 ms, mean = 0 ms, median = 0 ms, percentile(90%) = 0 ms, percentile(95%) = 0 ms, percentile(99%) = 0 ms
21:52:51.983: [03/12/2023-21:52:51] [I] Total Host Walltime: 3.03613 s
21:52:51.983: [03/12/2023-21:52:51] [I] Total GPU Compute Time: 3.03299 s
21:52:51.983: [03/12/2023-21:52:51] [I] Explanations of the performance metrics are printed in the verbose logs.
21:52:51.983: [03/12/2023-21:52:51] [I]
21:52:51.983: &&&& PASSED TensorRT.trtexec [TensorRT v8501] # C:/Program Files (x86)/SVP 4/rife\vsmlrt-cuda\trtexec --onnx=C:/Program Files (x86)/SVP 4/rife\models\rife_v2\rife_v4.4_ensemble.onnx --timingCacheFile=C:\Users\HTPC_U~1\AppData\Local\Temp\Program Files (x86)/SVP 4/rife\models\rife_v2\rife_v4.4_ensemble.onnx.1920x832_fp16_trt-8502_cudnn_I-fp16_O-fp16_NVIDIA-GeForce-RTX-3060-Ti_139fe355.engine.cache --device=0 --saveEngine=C:\Users\HTPC_U~1\AppData\Local\Temp\Program Files (x86)/SVP 4/rife\models\rife_v2\rife_v4.4_ensemble.onnx.1920x832_fp16_trt-8502_cudnn_I-fp16_O-fp16_NVIDIA-GeForce-RTX-3060-Ti_139fe355.engine --shapes=input:1x7x832x1920 --fp16 --tacticSources=-CUBLAS,-CUBLAS_LT --useCudaGraph --noDataTransfers --inputIOFormats=fp16:chw --outputIOFormats=fp16:chw
21:52:53.034: CUDA lazy loading is not enabled. Enabling it can significantly reduce device memory usage. See `CUDA_MODULE_LOADING` in https://docs.nvidia.com/cuda/c…guide/index.html#env-vars
21:52:53.095: CUDA lazy loading is not enabled. Enabling it can significantly reduce device memory usage. See `CUDA_MODULE_LOADING` in https://docs.nvidia.com/cuda/c…guide/index.html#env-vars
21:52:53.141: CUDA lazy loading is not enabled. Enabling it can significantly reduce device memory usage. See `CUDA_MODULE_LOADING` in https://docs.nvidia.com/cuda/c…guide/index.html#env-vars
21:52:53.188: CUDA lazy loading is not enabled. Enabling it can significantly reduce device memory usage. See `CUDA_MODULE_LOADING` in https://docs.nvidia.com/cuda/c…guide/index.html#env-vars
21:52:53.297: VO: [lavc] 1920x820 yuv420p
21:52:53.297: [vo/lavc] Opening encoder: NVIDIA NVENC hevc encoder [hevc_nvenc]
21:52:53.448: [encode] Opening output file: M:/Filme/Con Air - Extended Edition (1997)/Con Air - Extended Edition (1997) 1080p AC3.SVP.temporary.mkv
21:52:53.566: [encode] Opening muxer: Matroska [matroska]
21:52:53.619: V: 00:00:00 / 01:56:39 (0%) {1.0min 0.0MB} Cache: 514s/150MB
21:53:53.655: V: 00:01:00 / 01:56:39 (1%) {96.8min 59.8fps 11012.0MB} Cache: 543s/150MB
21:54:54.033: V: 00:01:59 / 01:56:39 (2%) {113.2min 59.4fps 12883.6MB} Cache: 541s/150MB
21:55:54.060: V: 00:02:56 / 01:56:39 (3%) {124.2min 58.6fps 13860.9MB} Cache: 522s/150MB
21:56:54.085: V: 00:03:54 / 01:56:39 (3%) {114.8min 58.4fps 13128.3MB} Cache: 541s/150MB
21:57:54.115: V: 00:04:53 / 01:56:39 (4%) {96.5min 58.5fps 11272.3MB} Cache: 596s/150MB
21:58:54.723: V: 00:05:54 / 01:56:39 (5%) {102.9min 58.8fps 11784.8MB} Cache: 579s/150MB
21:59:54.744: V: 00:06:52 / 01:56:39 (6%) {104.0min 58.8fps 12256.4MB} Cache: 578s/150MB
22:00:54.753: V: 00:07:52 / 01:56:39 (7%) {105.5min 58.9fps 12485.8MB} Cache: 576s/150MB
22:01:54.777: V: 00:08:51 / 01:56:39 (8%) {109.5min 58.9fps 12989.7MB} Cache: 549s/150MB
22:02:54.783: V: 00:09:50 / 01:56:39 (8%) {110.6min 58.9fps 13342.6MB} Cache: 535s/150MB
22:03:55.423: V: 00:10:48 / 01:56:39 (9%) {111.1min 58.8fps 13576.4MB} Cache: 539s/150MB
22:04:55.427: V: 00:11:49 / 01:56:39 (10%) {108.6min 58.9fps 13406.6MB} Cache: 567s/150MB
22:05:55.457: V: 00:12:48 / 01:56:39 (11%) {108.3min 58.9fps 13477.2MB} Cache: 567s/150MB
22:06:55.463: V: 00:13:48 / 01:56:39 (12%) {105.1min 59.0fps 13262.4MB} Cache: 574s/150MB
22:07:55.494: V: 00:14:50 / 01:56:39 (13%) {105.6min 59.2fps 13351.2MB} Cache: 538s/150MB
22:08:55.520: V: 00:15:48 / 01:56:39 (14%) {104.7min 59.2fps 13419.2MB} Cache: 516s/150MB
22:09:55.547: V: 00:16:49 / 01:56:39 (14%) {103.9min 59.3fps 13432.2MB} Cache: 505s/150MB
22:10:55.617: V: 00:17:46 / 01:56:39 (15%) {103.1min 59.1fps 13486.0MB} Cache: 508s/150MB
22:11:55.631: V: 00:18:45 / 01:56:39 (16%) {101.6min 59.1fps 13440.1MB} Cache: 514s/150MB
22:12:55.637: V: 00:19:45 / 01:56:39 (17%) {101.6min 59.2fps 13507.2MB} Cache: 515s/150MB
22:13:55.639: V: 00:20:44 / 01:56:39 (18%) {101.7min 59.1fps 13614.7MB} Cache: 506s/150MB
22:14:56.015: V: 00:21:44 / 01:56:39 (19%) {101.4min 59.2fps 13761.8MB} Cache: 506s/150MB
22:15:56.032: V: 00:22:43 / 01:56:39 (19%) {100.9min 59.2fps 13819.1MB} Cache: 516s/150MB
22:16:56.048: V: 00:23:47 / 01:56:39 (20%) {96.7min 59.4fps 13545.9MB} Cache: 572s/150MB
22:17:56.098: V: 00:24:51 / 01:56:39 (21%) {92.8min 59.6fps 13243.5MB} Cache: 613s/150MB
22:18:56.122: V: 00:25:51 / 01:56:39 (22%) {92.2min 59.6fps 13232.5MB} Cache: 611s/150MB
22:19:56.157: V: 00:26:50 / 01:56:39 (23%) {91.3min 59.5fps 13271.7MB} Cache: 624s/150MB
22:20:56.179: V: 00:27:51 / 01:56:39 (24%) {90.8min 59.6fps 13345.5MB} Cache: 612s/150MB
22:21:56.194: V: 00:28:49 / 01:56:39 (25%) {90.4min 59.6fps 13401.7MB} Cache: 602s/150MB
22:22:56.204: V: 00:29:49 / 01:56:39 (26%) {89.7min 59.6fps 13426.3MB} Cache: 611s/150MB
22:23:56.232: V: 00:30:47 / 01:56:39 (26%) {89.8min 59.5fps 13527.2MB} Cache: 606s/150MB
22:24:56.237: V: 00:31:43 / 01:56:39 (27%) {89.6min 59.4fps 13594.9MB} Cache: 608s/150MB
22:25:56.264: V: 00:32:40 / 01:56:39 (28%) {89.5min 59.3fps 13685.4MB} Cache: 594s/149MB
22:26:56.281: V: 00:33:40 / 01:56:39 (29%) {89.2min 59.3fps 13764.1MB} Cache: 577s/150MB
22:27:56.304: V: 00:34:39 / 01:56:39 (30%) {88.2min 59.3fps 13785.9MB} Cache: 581s/150MB
22:28:56.314: V: 00:35:39 / 01:56:39 (31%) {87.3min 59.3fps 13793.5MB} Cache: 579s/150MB
22:29:56.321: V: 00:36:38 / 01:56:39 (31%) {86.4min 59.4fps 13800.7MB} Cache: 590s/150MB
22:30:56.347: V: 00:37:38 / 01:56:39 (32%) {85.7min 59.3fps 13837.4MB} Cache: 593s/150MB
22:31:56.373: V: 00:38:37 / 01:56:39 (33%) {84.4min 59.3fps 13830.1MB} Cache: 608s/150MB
22:32:56.412: V: 00:39:35 / 01:56:39 (34%) {83.6min 59.3fps 13828.0MB} Cache: 617s/150MB
22:33:56.448: V: 00:40:35 / 01:56:39 (35%) {83.3min 59.3fps 13901.3MB} Cache: 604s/150MB
22:34:56.476: V: 00:41:36 / 01:56:39 (36%) {82.9min 59.4fps 13964.4MB} Cache: 612s/150MB
22:35:56.492: V: 00:42:36 / 01:56:39 (37%) {81.8min 59.4fps 13972.8MB} Cache: 627s/150MB
22:36:56.532: V: 00:43:36 / 01:56:39 (37%) {80.6min 59.4fps 13958.5MB} Cache: 623s/150MB
22:37:56.991: V: 00:44:37 / 01:56:39 (38%) {79.6min 59.4fps 13970.8MB} Cache: 618s/150MB
22:38:57.002: V: 00:45:38 / 01:56:39 (39%) {78.5min 59.4fps 13965.0MB} Cache: 628s/150MB
22:39:57.023: V: 00:46:39 / 01:56:39 (40%) {78.0min 59.5fps 14006.7MB} Cache: 614s/150MB
22:40:57.095: V: 00:47:38 / 01:56:39 (41%) {77.3min 59.5fps 14072.3MB} Cache: 604s/150MB
22:41:57.102: V: 00:48:38 / 01:56:39 (42%) {76.5min 59.5fps 14085.8MB} Cache: 618s/150MB
22:42:57.122: V: 00:49:38 / 01:56:39 (43%) {75.8min 59.5fps 14104.8MB} Cache: 610s/150MB
22:43:57.150: V: 00:50:37 / 01:56:39 (43%) {75.0min 59.5fps 14128.1MB} Cache: 628s/150MB
22:44:57.180: V: 00:51:38 / 01:56:39 (44%) {74.7min 59.5fps 14190.6MB} Cache: 604s/150MB
22:45:57.192: V: 00:52:38 / 01:56:39 (45%) {74.0min 59.5fps 14269.5MB} Cache: 585s/150MB
22:46:57.195: V: 00:53:40 / 01:56:39 (46%) {72.9min 59.6fps 14274.1MB} Cache: 592s/150MB
22:47:57.219: V: 00:54:42 / 01:56:39 (47%) {71.5min 59.6fps 14235.8MB} Cache: 583s/150MB
22:48:57.273: V: 00:55:42 / 01:56:39 (48%) {70.8min 59.6fps 14243.8MB} Cache: 568s/150MB
22:49:57.293: V: 00:56:43 / 01:56:39 (49%) {69.7min 59.6fps 14239.7MB} Cache: 575s/150MB
22:50:57.298: V: 00:57:43 / 01:56:39 (49%) {68.6min 59.6fps 14246.2MB} Cache: 572s/150MB
22:51:57.313: V: 00:58:45 / 01:56:39 (50%) {68.0min 59.7fps 14268.3MB} Cache: 554s/150MB
22:52:57.316: V: 00:59:40 / 01:56:39 (51%) {67.1min 59.6fps 14293.7MB} Cache: 536s/150MB
22:53:57.325: V: 01:00:39 / 01:56:39 (52%) {66.8min 59.6fps 14340.4MB} Cache: 506s/150MB
22:54:57.344: V: 01:01:37 / 01:56:39 (53%) {65.9min 59.6fps 14372.0MB} Cache: 500s/150MB
22:55:57.370: V: 01:02:36 / 01:56:39 (54%) {64.5min 59.6fps 14331.5MB} Cache: 517s/150MB
22:56:57.393: V: 01:03:36 / 01:56:39 (55%) {63.6min 59.6fps 14334.6MB} Cache: 524s/150MB
22:57:57.420: V: 01:04:34 / 01:56:39 (55%) {62.0min 59.6fps 14291.4MB} Cache: 529s/150MB
22:58:57.637: V: 01:05:37 / 01:56:39 (56%) {61.0min 59.6fps 14257.7MB} Cache: 513s/150MB
22:59:57.652: V: 01:06:38 / 01:56:39 (57%) {60.0min 59.6fps 14265.6MB} Cache: 509s/150MB
23:00:57.657: V: 01:07:38 / 01:56:39 (58%) {58.8min 59.6fps 14265.3MB} Cache: 516s/149MB
23:01:57.684: V: 01:08:40 / 01:56:39 (59%) {57.0min 59.7fps 14194.8MB} Cache: 586s/150MB
23:02:57.726: V: 01:09:43 / 01:56:39 (60%) {55.7min 59.7fps 14154.9MB} Cache: 592s/150MB
23:03:58.698: V: 01:10:44 / 01:56:39 (61%) {54.6min 59.7fps 14157.6MB} Cache: 592s/150MB
23:04:58.723: V: 01:11:46 / 01:56:39 (62%) {53.8min 59.7fps 14165.8MB} Cache: 576s/150MB
23:05:58.744: V: 01:12:45 / 01:56:39 (62%) {53.0min 59.7fps 14202.5MB} Cache: 552s/150MB
23:06:58.751: V: 01:13:43 / 01:56:39 (63%) {51.5min 59.7fps 14152.8MB} Cache: 543s/150MB
23:07:59.028: V: 01:14:45 / 01:56:39 (64%) {50.2min 59.7fps 14122.5MB} Cache: 536s/150MB
23:08:59.041: V: 01:15:45 / 01:56:39 (65%) {49.0min 59.7fps 14096.5MB} Cache: 530s/149MB
23:09:59.043: V: 01:16:48 / 01:56:39 (66%) {48.2min 59.8fps 14107.0MB} Cache: 508s/150MB
23:10:59.118: V: 01:17:50 / 01:56:39 (67%) {47.9min 59.8fps 14159.1MB} Cache: 469s/150MB
23:11:59.145: V: 01:18:48 / 01:56:39 (68%) {47.4min 59.8fps 14232.1MB} Cache: 440s/150MB
23:12:59.221: V: 01:19:47 / 01:56:39 (68%) {46.3min 59.8fps 14234.1MB} Cache: 427s/150MB
23:13:59.258: V: 01:20:46 / 01:56:39 (69%) {45.3min 59.8fps 14239.5MB} Cache: 407s/150MB
23:14:59.768: V: 01:21:46 / 01:56:39 (70%) {43.6min 59.8fps 14187.8MB} Cache: 406s/150MB
23:15:59.794: V: 01:22:48 / 01:56:39 (71%) {41.8min 59.8fps 14097.2MB} Cache: 415s/150MB
23:16:59.821: V: 01:23:52 / 01:56:39 (72%) {40.3min 59.8fps 14055.4MB} Cache: 415s/150MB
23:17:59.821: V: 01:24:55 / 01:56:39 (73%) {38.9min 59.9fps 14012.1MB} Cache: 434s/150MB
23:18:59.847: V: 01:25:57 / 01:56:39 (74%) {37.4min 59.9fps 13962.7MB} Cache: 467s/150MB
23:19:59.847: V: 01:27:05 / 01:56:39 (75%) {35.6min 60.0fps 13887.7MB} Cache: 481s/150MB
23:20:59.945: V: 01:28:11 / 01:56:39 (76%) {33.7min 60.1fps 13814.8MB} Cache: 489s/150MB
23:21:59.957: V: 01:29:18 / 01:56:39 (77%) {32.1min 60.1fps 13757.9MB} Cache: 513s/150MB
23:22:59.976: V: 01:30:24 / 01:56:39 (77%) {30.7min 60.2fps 13718.2MB} Cache: 493s/149MB
23:24:00.015: V: 01:31:26 / 01:56:39 (78%) {29.5min 60.2fps 13704.1MB} Cache: 473s/150MB
23:25:00.016: V: 01:32:29 / 01:56:39 (79%) {28.7min 60.3fps 13715.7MB} Cache: 428s/150MB
23:26:00.054: V: 01:33:31 / 01:56:39 (80%) {27.8min 60.3fps 13744.9MB} Cache: 391s/150MB
23:27:00.141: V: 01:34:33 / 01:56:39 (81%) {26.7min 60.3fps 13739.3MB} Cache: 377s/150MB
23:28:00.145: V: 01:35:35 / 01:56:39 (82%) {25.3min 60.3fps 13707.2MB} Cache: 377s/150MB
23:29:00.163: V: 01:36:36 / 01:56:39 (83%) {24.0min 60.3fps 13673.6MB} Cache: 376s/150MB
23:30:00.190: V: 01:37:38 / 01:56:39 (84%) {23.0min 60.3fps 13670.0MB} Cache: 344s/150MB
23:31:00.218: V: 01:38:42 / 01:56:39 (85%) {21.3min 60.4fps 13612.5MB} Cache: 345s/150MB
23:32:00.225: V: 01:39:47 / 01:56:39 (86%) {18.9min 60.4fps 13469.1MB} Cache: 365s/150MB
23:33:00.532: V: 01:40:49 / 01:56:39 (86%) {17.3min 60.4fps 13393.0MB} Cache: 351s/150MB
23:34:00.822: V: 01:41:51 / 01:56:39 (87%) {16.0min 60.4fps 13352.8MB} Cache: 352s/150MB
23:35:00.829: V: 01:42:54 / 01:56:39 (88%) {14.7min 60.5fps 13321.3MB} Cache: 342s/150MB
23:36:00.830: V: 01:44:00 / 01:56:39 (89%) {12.8min 60.5fps 13234.4MB} Cache: 389s/150MB
23:37:00.838: V: 01:45:05 / 01:56:39 (90%) {11.5min 60.6fps 13202.0MB} Cache: 472s/150MB
23:38:00.939: V: 01:46:06 / 01:56:39 (91%) {9.4min 60.6fps 13099.2MB} Cache: 562s/150MB
23:39:00.957: V: 01:47:09 / 01:56:39 (92%) {7.7min 60.6fps 13020.8MB} Cache: 568s/135MB
23:40:00.976: V: 01:48:11 / 01:56:39 (93%) {6.5min 60.6fps 12983.2MB} Cache: 506s/112MB
23:41:01.008: V: 01:49:13 / 01:56:39 (94%) {5.2min 60.6fps 12958.4MB} Cache: 444s/90MB
23:42:01.055: V: 01:50:17 / 01:56:39 (95%) {4.3min 60.6fps 12953.3MB} Cache: 379s/73MB
23:43:01.063: V: 01:51:19 / 01:56:39 (95%) {3.9min 60.7fps 13003.1MB} Cache: 318s/66MB
23:44:01.154: V: 01:52:19 / 01:56:39 (96%) {3.2min 60.6fps 13060.9MB} Cache: 258s/54MB
23:45:01.196: V: 01:53:22 / 01:56:39 (97%) {2.5min 60.7fps 13116.2MB} Cache: 195s/43MB
23:46:01.208: V: 01:54:26 / 01:56:39 (98%) {1.6min 60.7fps 13151.3MB} Cache: 131s/28MB
23:47:01.234: V: 01:55:31 / 01:56:39 (99%) {0.8min 60.7fps 13167.5MB} Cache: 66s/12MB
23:48:01.288: [encode] video: encoded 13817356403 bytes
23:48:01.288: [encode] audio: encoded 0 bytes
23:48:01.288: [encode] muxing overhead 3334476 bytes
23:48:01.316: Exiting... (End of file)
23:48:01.509: ===== Starting mkvmerge ======
23:48:01.509: Command line: C:\Program Files (x86)\SVP 4\extensions\code\mkvmerge.exe -o M:/Filme/Con Air - Extended Edition (1997)/Con Air - Extended Edition (1997) 1080p AC3.SVP.mkv M:/Filme/Con Air - Extended Edition (1997)/Con Air - Extended Edition (1997) 1080p AC3.SVP.temporary.mkv -D M:/Filme/Con Air - Extended Edition (1997)/Con Air - Extended Edition (1997) 1080p AC3.mkv
23:48:01.582: mkvmerge v73.0.0 ('25 or 6 to 4') 64-bit
23:48:02.118: 'M:/Filme/Con Air - Extended Edition (1997)/Con Air - Extended Edition (1997) 1080p AC3.SVP.temporary.mkv': Using the demultiplexer for the format 'Matroska'.
23:48:02.274: 'M:/Filme/Con Air - Extended Edition (1997)/Con Air - Extended Edition (1997) 1080p AC3.mkv': Using the demultiplexer for the format 'Matroska'.
23:48:02.274: 'M:/Filme/Con Air - Extended Edition (1997)/Con Air - Extended Edition (1997) 1080p AC3.SVP.temporary.mkv' track 0: Using the output module for the format 'HEVC/H.265'.
23:48:02.274: 'M:/Filme/Con Air - Extended Edition (1997)/Con Air - Extended Edition (1997) 1080p AC3.mkv' track 1: Using the output module for the format 'AC-3'.
23:48:02.384: The file 'M:/Filme/Con Air - Extended Edition (1997)/Con Air - Extended Edition (1997) 1080p AC3.SVP.mkv' has been opened for writing.
23:48:02.509: Progress: 0%
23:48:11.459: Progress: 1%
23:48:21.123: Progress: 2%
23:48:31.443: Progress: 3%
23:48:44.197: Progress: 4%
23:48:55.845: Progress: 5%
23:49:07.268: Progress: 6%
23:49:19.645: Progress: 7%
23:49:29.831: Progress: 8%
23:49:43.886: Progress: 9%
23:49:58.256: Progress: 10%
23:50:09.906: Progress: 11%
23:50:22.860: Progress: 12%
23:50:35.453: Progress: 13%
23:50:46.850: Progress: 14%
23:50:59.396: Progress: 15%
23:51:13.577: Progress: 16%
23:51:26.512: Progress: 17%
23:51:39.593: Progress: 18%
23:51:57.719: Progress: 19%
23:52:14.809: Progress: 20%
23:52:28.884: Progress: 21%
23:52:42.189: Progress: 22%
23:53:00.232: Progress: 23%
23:53:14.150: Progress: 24%
23:53:29.625: Progress: 25%
23:53:43.911: Progress: 26%
23:53:57.053: Progress: 27%
23:54:10.572: Progress: 28%
23:54:21.394: Progress: 29%
23:54:36.839: Progress: 30%
23:54:50.502: Progress: 31%
23:55:04.388: Progress: 32%
23:55:16.821: Progress: 33%
23:55:29.264: Progress: 34%
23:55:42.760: Progress: 35%
23:55:57.362: Progress: 36%
23:56:08.725: Progress: 37%
23:56:20.713: Progress: 38%
23:56:35.083: Progress: 39%
23:56:47.851: Progress: 40%
23:57:02.324: Progress: 41%
23:57:14.845: Progress: 42%
23:57:27.842: Progress: 43%
23:57:41.694: Progress: 44%
23:57:55.860: Progress: 45%
23:58:09.823: Progress: 46%
23:58:18.854: Progress: 47%
23:58:31.831: Progress: 48%
23:58:46.268: Progress: 49%
23:59:00.742: Progress: 50%
23:59:15.445: Progress: 51%
23:59:28.075: Progress: 52%
23:59:45.071: Progress: 53%
23:59:57.856: Progress: 54%
00:00:16.825: Progress: 55%
00:00:33.568: Progress: 56%
00:00:48.369: Progress: 57%
00:01:02.835: Progress: 58%
00:01:18.738: Progress: 59%
00:01:31.217: Progress: 60%
00:01:50.936: Progress: 61%
00:02:10.297: Progress: 62%
00:02:27.589: Progress: 63%
00:02:44.425: Progress: 64%
00:03:03.123: Progress: 65%
00:03:15.540: Progress: 66%
00:03:31.160: Progress: 67%
00:03:48.237: Progress: 68%
00:04:04.384: Progress: 69%
00:04:24.539: Progress: 70%
00:04:39.564: Progress: 71%
00:04:57.140: Progress: 72%
00:05:12.392: Progress: 73%
00:05:29.163: Progress: 74%
00:05:47.732: Progress: 75%
00:06:02.960: Progress: 76%
00:06:21.723: Progress: 77%
00:06:34.868: Progress: 78%
00:06:47.707: Progress: 79%
00:07:06.375: Progress: 80%
00:07:22.638: Progress: 81%
00:07:40.581: Progress: 82%
00:07:54.350: Progress: 83%
00:08:09.152: Progress: 84%
00:08:25.597: Progress: 85%
00:08:35.404: Progress: 86%
00:08:57.243: Progress: 87%
00:09:14.902: Progress: 88%
00:09:32.612: Progress: 89%
00:09:45.712: Progress: 90%
00:09:59.446: Progress: 91%
00:10:17.409: Progress: 92%
00:10:33.860: Progress: 93%
00:10:49.514: Progress: 94%
00:11:08.338: Progress: 95%
00:11:23.123: Progress: 96%
00:11:36.312: Progress: 97%
00:11:48.711: Progress: 98%
00:12:03.839: Progress: 99%
00:12:15.965: Progress: 100%
00:12:15.965: The cue entries (the index) are being written...
00:12:24.171: Multiplexing took 24 minutes 22 seconds.
-
-
Wenn ich einen FullHD Film (Avengers) abspiele, kann ich auf RIFE AI umschalten sofern ich bei Avisynth bleibe und "an Bildschirm" einstelle.
Es startet dann allerdings keine CMD. Gehe ich auf x2 oder x2.5 stottert es.
Stelle ich auf VPS um, so öffnet sich zwar die CMD, aber die bleibt leer (5 Min gewartet) während im Hintergrund nur der Ton läuft.
-
-
Meine Fresse bin ich ein Trottel
Ich habe die vsmirt.py per Rechtsklick aus der File Übersicht in GitHub gespeichert und auch brav
die Bool Werte editiert.
Eben fiel mir das auf, dass ganz oben DOCTYPE steht.
Also anständig denn Code per Copy&Paste in die Datei kopiert und die Bool Werte erneut angepasst,
jetzt macht der Textblock auch mehr Sinn für ein Python Skript.
Im MPV funktioniert es jetzt bedingt, er zeigt beim Wechsel auf RIFE AI die SVP Meldung an, allerdings mit Farbfehler (s/w).
Außerdem kommt die CMD mit den richtigen Models hoch.
Zitat&&&& RUNNING TensorRT.trtexec [TensorRT v8501] # C:/Program Files (x86)/SVP 4/rife\vsmlrt-cuda\trtexec --onnx=C:/Program Files (x86)/SVP 4/rife\models\rife_v2\rife_v4.6_ensemble.onnx --timingCacheFile=C:\Users\HTPC_U~1\AppData\Local\Temp\Program Files (x86)/SVP 4/rife\models\rife_v2\rife_v4.6_ensemble.onnx.3840x2144_fp16_trt-8502_cudnn_I-fp16_O-fp16_NVIDIA-GeForce-RTX-3060-Ti_f0d2d789.engine.cache --device=0 --saveEngine=C:\Users\HTPC_U~1\AppData\Local\Temp\Program Files (x86)/SVP 4/rife\models\rife_v2\rife_v4.6_ensemble.onnx.3840x2144_fp16_trt-8502_cudnn_I-fp16_O-fp16_NVIDIA-GeForce-RTX-3060-Ti_f0d2d789.engine --shapes=input:1x7x2144x3840 --fp16 --tacticSources=-CUBLAS,-CUBLAS_LT --useCudaGraph --noDataTransfers --inputIOFormats=fp16:chw --outputIOFormats=fp16:chw
[03/12/2023-15:50:25] [I] === Model Options ===
[03/12/2023-15:50:25] [I] Format: ONNX
[03/12/2023-15:50:25] [I] Model: C:/Program Files (x86)/SVP 4/rife\models\rife_v2\rife_v4.6_ensemble.onnx
[03/12/2023-15:50:25] [I] Output:
[03/12/2023-15:50:25] [I] === Build Options ===
[03/12/2023-15:50:25] [I] Max batch: explicit batch
[03/12/2023-15:50:25] [I] Memory Pools: workspace: default, dlaSRAM: default, dlaLocalDRAM: default, dlaGlobalDRAM: default
[03/12/2023-15:50:25] [I] minTiming: 1
[03/12/2023-15:50:25] [I] avgTiming: 8
[03/12/2023-15:50:25] [I] Precision: FP32+FP16
[03/12/2023-15:50:25] [I] LayerPrecisions:
[03/12/2023-15:50:25] [I] Calibration:
[03/12/2023-15:50:25] [I] Refit: Disabled
[03/12/2023-15:50:25] [I] Sparsity: Disabled
[03/12/2023-15:50:25] [I] Safe mode: Disabled
[03/12/2023-15:50:25] [I] DirectIO mode: Disabled
[03/12/2023-15:50:25] [I] Restricted mode: Disabled
[03/12/2023-15:50:25] [I] Build only: Disabled
[03/12/2023-15:50:25] [I] Save engine: C:\Users\HTPC_U~1\AppData\Local\Temp\Program Files (x86)/SVP 4/rife\models\rife_v2\rife_v4.6_ensemble.onnx.3840x2144_fp16_trt-8502_cudnn_I-fp16_O-fp16_NVIDIA-GeForce-RTX-3060-Ti_f0d2d789.engine
[03/12/2023-15:50:25] [I] Load engine:
[03/12/2023-15:50:25] [I] Profiling verbosity: 0
[03/12/2023-15:50:25] [I] Tactic sources: cublas [OFF], cublasLt [OFF],
[03/12/2023-15:50:25] [I] timingCacheMode: global
[03/12/2023-15:50:25] [I] timingCacheFile: C:\Users\HTPC_U~1\AppData\Local\Temp\Program Files (x86)/SVP 4/rife\models\rife_v2\rife_v4.6_ensemble.onnx.3840x2144_fp16_trt-8502_cudnn_I-fp16_O-fp16_NVIDIA-GeForce-RTX-3060-Ti_f0d2d789.engine.cache
[03/12/2023-15:50:25] [I] Heuristic: Disabled
[03/12/2023-15:50:25] [I] Preview Features: Use default preview flags.
[03/12/2023-15:50:25] [I] Input(s): fp16:chw
[03/12/2023-15:50:25] [I] Output(s): fp16:chw
[03/12/2023-15:50:25] [I] Input build shape: input=1x7x2144x3840+1x7x2144x3840+1x7x2144x3840
[03/12/2023-15:50:25] [I] Input calibration shapes: model
[03/12/2023-15:50:25] [I] === System Options ===
[03/12/2023-15:50:25] [I] Device: 0
[03/12/2023-15:50:25] [I] DLACore:
[03/12/2023-15:50:25] [I] Plugins:
[03/12/2023-15:50:25] [I] === Inference Options ===
[03/12/2023-15:50:25] [I] Batch: Explicit
[03/12/2023-15:50:25] [I] Input inference shape: input=1x7x2144x3840
[03/12/2023-15:50:25] [I] Iterations: 10
[03/12/2023-15:50:25] [I] Duration: 3s (+ 200ms warm up)
[03/12/2023-15:50:25] [I] Sleep time: 0ms
[03/12/2023-15:50:25] [I] Idle time: 0ms
[03/12/2023-15:50:25] [I] Streams: 1
[03/12/2023-15:50:25] [I] ExposeDMA: Disabled
[03/12/2023-15:50:25] [I] Data transfers: Disabled
[03/12/2023-15:50:25] [I] Spin-wait: Disabled
[03/12/2023-15:50:25] [I] Multithreading: Disabled
[03/12/2023-15:50:25] [I] CUDA Graph: Enabled
[03/12/2023-15:50:25] [I] Separate profiling: Disabled
[03/12/2023-15:50:25] [I] Time Deserialize: Disabled
[03/12/2023-15:50:25] [I] Time Refit: Disabled
[03/12/2023-15:50:25] [I] NVTX verbosity: 0
[03/12/2023-15:50:25] [I] Persistent Cache Ratio: 0
[03/12/2023-15:50:25] [I] Inputs:
[03/12/2023-15:50:25] [I] === Reporting Options ===
[03/12/2023-15:50:25] [I] Verbose: Disabled
[03/12/2023-15:50:25] [I] Averages: 10 inferences
[03/12/2023-15:50:25] [I] Percentiles: 90,95,99
[03/12/2023-15:50:25] [I] Dump refittable layers:Disabled
[03/12/2023-15:50:25] [I] Dump output: Disabled
[03/12/2023-15:50:25] [I] Profile: Disabled
[03/12/2023-15:50:25] [I] Export timing to JSON file:
[03/12/2023-15:50:25] [I] Export output to JSON file:
[03/12/2023-15:50:25] [I] Export profile to JSON file:
[03/12/2023-15:50:25] [I]
[03/12/2023-15:50:25] [I] === Device Information ===
[03/12/2023-15:50:25] [I] Selected Device: NVIDIA GeForce RTX 3060 Ti
[03/12/2023-15:50:25] [I] Compute Capability: 8.6
[03/12/2023-15:50:25] [I] SMs: 38
[03/12/2023-15:50:25] [I] Compute Clock Rate: 1.665 GHz
[03/12/2023-15:50:25] [I] Device Global Memory: 8191 MiB
[03/12/2023-15:50:25] [I] Shared Memory per SM: 100 KiB
[03/12/2023-15:50:25] [I] Memory Bus Width: 256 bits (ECC disabled)
[03/12/2023-15:50:25] [I] Memory Clock Rate: 7.001 GHz
[03/12/2023-15:50:25] [I]
[03/12/2023-15:50:25] [I] TensorRT version: 8.5.1
[03/12/2023-15:50:26] [I] [TRT] [MemUsageChange] Init CUDA: CPU +449, GPU +0, now: CPU 10062, GPU 1183 (MiB)
[03/12/2023-15:50:28] [I] [TRT] [MemUsageChange] Init builder kernel library: CPU +423, GPU +118, now: CPU 10968, GPU 1301 (MiB)
[03/12/2023-15:50:28] [W] [TRT] CUDA lazy loading is not enabled. Enabling it can significantly reduce device memory usage. See `CUDA_MODULE_LOADING` in https://docs.nvidia.com/cuda/c…guide/index.html#env-vars
[03/12/2023-15:50:28] [I] Start parsing network model
[03/12/2023-15:50:28] [I] [TRT] ----------------------------------------------------------------
[03/12/2023-15:50:28] [I] [TRT] Input filename: C:/Program Files (x86)/SVP 4/rife\models\rife_v2\rife_v4.6_ensemble.onnx
[03/12/2023-15:50:28] [I] [TRT] ONNX IR version: 0.0.8
[03/12/2023-15:50:28] [I] [TRT] Opset version: 16
[03/12/2023-15:50:28] [I] [TRT] Producer name: pytorch
[03/12/2023-15:50:28] [I] [TRT] Producer version: 2.0.0
[03/12/2023-15:50:28] [I] [TRT] Domain:
[03/12/2023-15:50:28] [I] [TRT] Model version: 0
[03/12/2023-15:50:28] [I] [TRT] Doc string:
[03/12/2023-15:50:28] [I] [TRT] ----------------------------------------------------------------
[03/12/2023-15:50:28] [W] [TRT] onnx2trt_utils.cpp:377: Your ONNX model has been generated with INT64 weights, while TensorRT does not natively support INT64. Attempting to cast down to INT32.
[03/12/2023-15:50:28] [W] [TRT] onnx2trt_utils.cpp:403: One or more weights outside the range of INT32 was clamped
[03/12/2023-15:50:28] [W] [TRT] onnx2trt_utils.cpp:403: One or more weights outside the range of INT32 was clamped
[03/12/2023-15:50:28] [W] [TRT] onnx2trt_utils.cpp:403: One or more weights outside the range of INT32 was clamped
[03/12/2023-15:50:28] [W] [TRT] onnx2trt_utils.cpp:403: One or more weights outside the range of INT32 was clamped
[03/12/2023-15:50:29] [W] [TRT] onnx2trt_utils.cpp:403: One or more weights outside the range of INT32 was clamped
[03/12/2023-15:50:29] [I] Finish parsing network model
[03/12/2023-15:50:29] [W] Could not read timing cache from: C:\Users\HTPC_U~1\AppData\Local\Temp\Program Files (x86)/SVP 4/rife\models\rife_v2\rife_v4.6_ensemble.onnx.3840x2144_fp16_trt-8502_cudnn_I-fp16_O-fp16_NVIDIA-GeForce-RTX-3060-Ti_f0d2d789.engine.cache. A new timing cache will be generated and written.
[03/12/2023-15:50:31] [I] [TRT] [MemUsageChange] Init cuDNN: CPU +1090, GPU +406, now: CPU 11762, GPU 1707 (MiB)
[03/12/2023-15:50:31] [I] [TRT] Global timing cache in use. Profiling results in this builder pass will be stored.
[03/12/2023-15:58:33] [I] [TRT] Total Activation Memory: 15198938624
[03/12/2023-15:59:49] [I] [TRT] Detected 4 inputs and 1 output network tensors.
[03/12/2023-15:59:49] [I] [TRT] Total Host Persistent Memory: 169056
[03/12/2023-15:59:49] [I] [TRT] Total Device Persistent Memory: 97262592
[03/12/2023-15:59:49] [I] [TRT] Total Scratch Memory: 98795520
[03/12/2023-15:59:49] [I] [TRT] [MemUsageStats] Peak memory usage of TRT CPU/GPU memory allocators: CPU 56 MiB, GPU 4736 MiB
[03/12/2023-15:59:49] [I] [TRT] [BlockAssignment] Started assigning block shifts. This will take 168 steps to complete.
[03/12/2023-15:59:49] [I] [TRT] [BlockAssignment] Algorithm ShiftNTopDown took 15.1665ms to assign 13 blocks to 168 nodes requiring 1202013184 bytes.
[03/12/2023-15:59:49] [I] [TRT] Total Activation Memory: 1202013184
[03/12/2023-15:59:50] [W] [TRT] TensorRT encountered issues when converting weights between types and that could affect accuracy.
[03/12/2023-15:59:50] [W] [TRT] If this is not the desired behavior, please modify the weights or retrain with regularization to adjust the magnitude of the weights.
[03/12/2023-15:59:50] [W] [TRT] Check verbose logs for the list of affected weights.
[03/12/2023-15:59:50] [W] [TRT] - 88 weights are affected by this issue: Detected subnormal FP16 values.
[03/12/2023-15:59:50] [W] [TRT] - 16 weights are affected by this issue: Detected values less than smallest positive FP16 subnormal value and converted them to the FP16 minimum subnormalized value.
[03/12/2023-15:59:50] [I] [TRT] [MemUsageChange] TensorRT-managed allocation in building engine: CPU +20, GPU +103, now: CPU 20, GPU 103 (MiB)
[03/12/2023-15:59:50] [I] [TRT] [MemUsageChange] Init CUDA: CPU +0, GPU +0, now: CPU 13380, GPU 1933 (MiB)
[03/12/2023-15:59:50] [W] [TRT] CUDA lazy loading is not enabled. Enabling it can significantly reduce device memory usage. See `CUDA_MODULE_LOADING` in https://docs.nvidia.com/cuda/c…guide/index.html#env-vars
[03/12/2023-15:59:50] [I] Saved 274240 bytes of timing cache to C:\Users\HTPC_U~1\AppData\Local\Temp\Program Files (x86)/SVP 4/rife\models\rife_v2\rife_v4.6_ensemble.onnx.3840x2144_fp16_trt-8502_cudnn_I-fp16_O-fp16_NVIDIA-GeForce-RTX-3060-Ti_f0d2d789.engine.cache
[03/12/2023-15:59:50] [I] Engine built in 564.417 sec.
[03/12/2023-15:59:50] [I] [TRT] Loaded engine size: 10 MiB
[03/12/2023-15:59:50] [I] [TRT] [MemUsageChange] TensorRT-managed allocation in engine deserialization: CPU +0, GPU +102, now: CPU 0, GPU 102 (MiB)
[03/12/2023-15:59:50] [I] Engine deserialized in 0.0180156 sec.
[03/12/2023-15:59:50] [I] [TRT] [MemUsageChange] TensorRT-managed allocation in IExecutionContext creation: CPU +0, GPU +1239, now: CPU 0, GPU 1341 (MiB)
[03/12/2023-15:59:50] [W] [TRT] CUDA lazy loading is not enabled. Enabling it can significantly reduce device memory usage. See `CUDA_MODULE_LOADING` in https://docs.nvidia.com/cuda/c…guide/index.html#env-vars
[03/12/2023-15:59:50] [I] Setting persistentCacheLimit to 0 bytes.
[03/12/2023-15:59:50] [I] Using random values for input input
[03/12/2023-15:59:51] [I] Created input binding for input with dimensions 1x7x2144x3840
[03/12/2023-15:59:51] [I] Using random values for output output
[03/12/2023-15:59:51] [I] Created output binding for output with dimensions 1x3x2144x3840
[03/12/2023-15:59:51] [I] Starting inference
Nachdem das beendet ist (2x knapp 10min bei The Hateful 8 UHD) läuft der Film an, aber der Cache ist mit 32secs winzig und nach 40secs bleibt
das Bild stehen, weil der Cache leer ist (lt. SVP ist die GPU zwischen 60 und 70%).
Wohlgemerkt bei RIFE AI 4.6 und den folgenden Settings:
Switche ich auf AI Model 4.4, so dauert die CMD fast 25 Minuten. Starte ich die Wiedergabe des gleichen Films erneut, beginnt die CMD wieder zu laufen,
ich dachte die Ergebnisse würden gespeichert werden, und man muss es nicht jedes Mal machen ?
Wie auch immer, hier das gleiche Ergebnis, ein winziger Cache läuft leer, dann hängt es für 1-2 Sekunden und läuft stotternd weiter bei 67% GPU Last.
Ich glaube mittlerweile, dass mein i3 -12100 hier die Limitierung ist und nicht die 3060ti.
Schalte ich auf automatisch kann ich den Film mit fix 60fps ohne Probleme laufen lassen im mpv.
Ebenso in MPC-HC.
Rife-AI funktioniert hingegen nicht beim MPC-HC (weder mit Vaporsynth noch mit Avisynth) der Player crashed nach max. 5 Sekunden ohne Fehlermeldung.
Scheint als ob ich das Projekt RIFE zurückstellen muss, bis ich mir irgendwann mal nen neuen HTPC baue. Das ist aber nicht in Sicht.
Meiner ist ja erst wenige Monate als und für MadVR absolut ausreichend.
Danke für eure Hilfe. Wie immer tolle Hilfe hier im Forum
-
-
Bei MPV kommt keinerlei Meldung nach dem Wechsel auf RIFE AI.
In MPV bekomme ich allerdings auch keine SVP Anzeige was die Bildrate angeht.
Da ich MPV noch nie wirklich genutzt habe, weiß ich nicht ob das normal ist. Einstellen kann ich da nix (Rechtsklick bzw. ALT bringt keinerlei Menüs).
-
-
Das ist eine gute Frage, ich glaube nein, aber finde den Fehler nicht.
Habe SVP neu installiert mit folgenden Optionen:
Dann habe ich meinen madVR 165 Ordner Inhalt nach "C:\Program Files (x86)\SVP 4\madVR " kopiert und anschließend,
der Anleitung folgend Avisynth 3.7.2 manuell.
Das RIFE AI Profil ist da und TensorRT ausgewählt
Im MPC-HC habe ich VPS eingestellt als prefered:
Was ich nicht im Menü finde sind die Variablen für VPS, bei mir will er nur VLC und MPV setzen:
Starte ich nun einen Film im MPC-HC, standsarmäßig automatic (wo kann man das denn ändern ?).
Wechsle ich im SVP manuell auf RIFE AI, so bekomme ich eine dauerhafte Fehlermeldung ins Bild und der FIlm läuft ohne eine Art der Vorberechnung:
Habe wie erwähnt nochmal von Vorne angefangen, die V2 Models sind aktuell nicht auf dem HTPC.
Bin für jede Hilfe dankbar
EDIT: Stelle ich RIFE 4.4 ein, so öffnet sich die CMD und ich warte brav 3 Minuten beginnt der Film, aber ebenfalls mit der Fehlermeldung.
Code
Alles anzeigen&&&& RUNNING TensorRT.trtexec [TensorRT v8501] # C:/Program Files (x86)/SVP 4/rife\vsmlrt-cuda\trtexec --onnx=C:/Program Files (x86)/SVP 4/rife\models\rife\rife_v4.4.onnx --timingCacheFile=C:\Users\HTPC_User\AppData\Roaming\SVP4\cache\Program Files (x86)/SVP 4/rife\models\rife\rife_v4.4.onnx.1920x800_fp16_trt-8502_cudnn_I-fp16_O-fp16_NVIDIA-GeForce-RTX-3060-Ti_a8b3b7a9.engine.cache --device=0 --saveEngine=C:\Users\HTPC_User\AppData\Roaming\SVP4\cache\Program Files (x86)/SVP 4/rife\models\rife\rife_v4.4.onnx.1920x800_fp16_trt-8502_cudnn_I-fp16_O-fp16_NVIDIA-GeForce-RTX-3060-Ti_a8b3b7a9.engine --shapes=input:1x11x800x1920 --fp16 --tacticSources=-CUBLAS,-CUBLAS_LT --useCudaGraph --noDataTransfers --inputIOFormats=fp16:chw --outputIOFormats=fp16:chw [03/12/2023-13:59:09] [I] === Model Options === [03/12/2023-13:59:09] [I] Format: ONNX [03/12/2023-13:59:09] [I] Model: C:/Program Files (x86)/SVP 4/rife\models\rife\rife_v4.4.onnx [03/12/2023-13:59:09] [I] Output: [03/12/2023-13:59:09] [I] === Build Options === [03/12/2023-13:59:09] [I] Max batch: explicit batch [03/12/2023-13:59:09] [I] Memory Pools: workspace: default, dlaSRAM: default, dlaLocalDRAM: default, dlaGlobalDRAM: default [03/12/2023-13:59:09] [I] minTiming: 1 [03/12/2023-13:59:09] [I] avgTiming: 8 [03/12/2023-13:59:09] [I] Precision: FP32+FP16 [03/12/2023-13:59:09] [I] LayerPrecisions: [03/12/2023-13:59:09] [I] Calibration: [03/12/2023-13:59:09] [I] Refit: Disabled [03/12/2023-13:59:09] [I] Sparsity: Disabled [03/12/2023-13:59:09] [I] Safe mode: Disabled [03/12/2023-13:59:09] [I] DirectIO mode: Disabled [03/12/2023-13:59:09] [I] Restricted mode: Disabled [03/12/2023-13:59:09] [I] Build only: Disabled [03/12/2023-13:59:09] [I] Save engine: C:\Users\HTPC_User\AppData\Roaming\SVP4\cache\Program Files (x86)/SVP 4/rife\models\rife\rife_v4.4.onnx.1920x800_fp16_trt-8502_cudnn_I-fp16_O-fp16_NVIDIA-GeForce-RTX-3060-Ti_a8b3b7a9.engine [03/12/2023-13:59:09] [I] Load engine: [03/12/2023-13:59:09] [I] Profiling verbosity: 0 [03/12/2023-13:59:09] [I] Tactic sources: cublas [OFF], cublasLt [OFF], [03/12/2023-13:59:09] [I] timingCacheMode: global [03/12/2023-13:59:09] [I] timingCacheFile: C:\Users\HTPC_User\AppData\Roaming\SVP4\cache\Program Files (x86)/SVP 4/rife\models\rife\rife_v4.4.onnx.1920x800_fp16_trt-8502_cudnn_I-fp16_O-fp16_NVIDIA-GeForce-RTX-3060-Ti_a8b3b7a9.engine.cache [03/12/2023-13:59:09] [I] Heuristic: Disabled [03/12/2023-13:59:09] [I] Preview Features: Use default preview flags. [03/12/2023-13:59:09] [I] Input(s): fp16:chw [03/12/2023-13:59:09] [I] Output(s): fp16:chw [03/12/2023-13:59:09] [I] Input build shape: input=1x11x800x1920+1x11x800x1920+1x11x800x1920 [03/12/2023-13:59:09] [I] Input calibration shapes: model [03/12/2023-13:59:09] [I] === System Options === [03/12/2023-13:59:09] [I] Device: 0 [03/12/2023-13:59:09] [I] DLACore: [03/12/2023-13:59:09] [I] Plugins: [03/12/2023-13:59:09] [I] === Inference Options === [03/12/2023-13:59:09] [I] Batch: Explicit [03/12/2023-13:59:09] [I] Input inference shape: input=1x11x800x1920 [03/12/2023-13:59:09] [I] Iterations: 10 [03/12/2023-13:59:09] [I] Duration: 3s (+ 200ms warm up) [03/12/2023-13:59:09] [I] Sleep time: 0ms [03/12/2023-13:59:09] [I] Idle time: 0ms [03/12/2023-13:59:09] [I] Streams: 1 [03/12/2023-13:59:09] [I] ExposeDMA: Disabled [03/12/2023-13:59:09] [I] Data transfers: Disabled [03/12/2023-13:59:09] [I] Spin-wait: Disabled [03/12/2023-13:59:09] [I] Multithreading: Disabled [03/12/2023-13:59:09] [I] CUDA Graph: Enabled [03/12/2023-13:59:09] [I] Separate profiling: Disabled [03/12/2023-13:59:09] [I] Time Deserialize: Disabled [03/12/2023-13:59:09] [I] Time Refit: Disabled [03/12/2023-13:59:09] [I] NVTX verbosity: 0 [03/12/2023-13:59:09] [I] Persistent Cache Ratio: 0 [03/12/2023-13:59:09] [I] Inputs: [03/12/2023-13:59:09] [I] === Reporting Options === [03/12/2023-13:59:09] [I] Verbose: Disabled [03/12/2023-13:59:09] [I] Averages: 10 inferences [03/12/2023-13:59:09] [I] Percentiles: 90,95,99 [03/12/2023-13:59:09] [I] Dump refittable layers:Disabled [03/12/2023-13:59:09] [I] Dump output: Disabled [03/12/2023-13:59:09] [I] Profile: Disabled [03/12/2023-13:59:09] [I] Export timing to JSON file: [03/12/2023-13:59:09] [I] Export output to JSON file: [03/12/2023-13:59:09] [I] Export profile to JSON file: [03/12/2023-13:59:09] [I] [03/12/2023-13:59:09] [I] === Device Information === [03/12/2023-13:59:09] [I] Selected Device: NVIDIA GeForce RTX 3060 Ti [03/12/2023-13:59:09] [I] Compute Capability: 8.6 [03/12/2023-13:59:09] [I] SMs: 38 [03/12/2023-13:59:09] [I] Compute Clock Rate: 1.665 GHz [03/12/2023-13:59:09] [I] Device Global Memory: 8191 MiB [03/12/2023-13:59:09] [I] Shared Memory per SM: 100 KiB [03/12/2023-13:59:09] [I] Memory Bus Width: 256 bits (ECC disabled) [03/12/2023-13:59:09] [I] Memory Clock Rate: 7.001 GHz [03/12/2023-13:59:09] [I] [03/12/2023-13:59:09] [I] TensorRT version: 8.5.1 [03/12/2023-13:59:10] [I] [TRT] [MemUsageChange] Init CUDA: CPU +413, GPU +0, now: CPU 11023, GPU 1183 (MiB) [03/12/2023-13:59:11] [I] [TRT] [MemUsageChange] Init builder kernel library: CPU +480, GPU +118, now: CPU 11926, GPU 1301 (MiB) [03/12/2023-13:59:11] [W] [TRT] CUDA lazy loading is not enabled. Enabling it can significantly reduce device memory usage. See `CUDA_MODULE_LOADING` in https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html#env-vars [03/12/2023-13:59:11] [I] Start parsing network model [03/12/2023-13:59:11] [I] [TRT] ---------------------------------------------------------------- [03/12/2023-13:59:11] [I] [TRT] Input filename: C:/Program Files (x86)/SVP 4/rife\models\rife\rife_v4.4.onnx [03/12/2023-13:59:11] [I] [TRT] ONNX IR version: 0.0.8 [03/12/2023-13:59:11] [I] [TRT] Opset version: 16 [03/12/2023-13:59:11] [I] [TRT] Producer name: pytorch [03/12/2023-13:59:11] [I] [TRT] Producer version: 1.12.0 [03/12/2023-13:59:11] [I] [TRT] Domain: [03/12/2023-13:59:11] [I] [TRT] Model version: 0 [03/12/2023-13:59:11] [I] [TRT] Doc string: [03/12/2023-13:59:11] [I] [TRT] ---------------------------------------------------------------- [03/12/2023-13:59:11] [W] [TRT] onnx2trt_utils.cpp:377: Your ONNX model has been generated with INT64 weights, while TensorRT does not natively support INT64. Attempting to cast down to INT32. [03/12/2023-13:59:11] [I] Finish parsing network model [03/12/2023-13:59:11] [W] Could not read timing cache from: C:\Users\HTPC_User\AppData\Roaming\SVP4\cache\Program Files (x86)/SVP 4/rife\models\rife\rife_v4.4.onnx.1920x800_fp16_trt-8502_cudnn_I-fp16_O-fp16_NVIDIA-GeForce-RTX-3060-Ti_a8b3b7a9.engine.cache. A new timing cache will be generated and written. [03/12/2023-13:59:12] [I] [TRT] [MemUsageChange] Init cuDNN: CPU +1109, GPU +406, now: CPU 12710, GPU 1707 (MiB) [03/12/2023-13:59:12] [I] [TRT] Global timing cache in use. Profiling results in this builder pass will be stored.
-
Bei 3x dachte ich erstmal an eine Art Stress Test, ich sehe ehrlich gesagt keinen Unterschied zwischen 2,5x und 3x.
Aber irgendwie kommt meine 3060ti dabei nichts ins Schwitzen, so um die 14% Last zeigt mir SVP an, und das wundert mich doch
bei RIFE AI mit allen Settings auf MAX.
-
Hab ich das richtig gemacht ?
unpack "rife_v2" into SVP 4\rife\models (i.e. there must be SVP 4\rife\models\rife_v2\rife_v4.6.onnx) - CHECK
replace SVP 4\rife\vsmlrt.py - CHECK
SVP neugestartet und Anwendungseinstellungen auf:
Bedeutet doch das Profil RIFE AI (3x) soll bei allen Filmen die unter 2160 sind und mit mpv (Player) abgespielt werden aktiv sein, oder ?
-
Sehe ich wie Viktor, egal welches Gerät du einschickst, steht immer irgendwas von „Alle Einstellungen können verloren gehen.“
-
Sorry, das geht auf mich. Da habe ich dich falsch verstanden.
Ich nutze es andersherum.
Anfang in iOS App bzw. WebApp und dann im Theater.
So herum funktioniert es.
-
Für das was du vorhast, kannst du ja mal Emby probieren.
Die haben einen eigenen Server und eine eigene "Theater App", ich weiß nicht ob die Theater App
mit einem Jellyfin Server funktioniert, aber genaugenommen ist Jellyfin ja ein Emby Fork, sollte theoretisch gehen.
Ich habe den Server zentral (NAS), die Theater App auf dem HTPC (dort kann man ext. Player definieren)
in Verbindung mit MPC+madvr. Und für unterwegs gibt's auch mobile Apps.
In der Konstellation funktioniert bei mir auch das merken der Wiedergabe Position in Verbindung mit MPC+madvr,
was so über KODI nicht funktioniert, wenn man einen anderen Player als den internen nutzt.
Und ja Emby benötigt für einiges eine Lizenz, aber ob du die für deine Anforderungen benötigst, oder ob
die Free Version reicht, musst du recherchieren.
-
Nachdem ich jetzt auf madVR gehe...
Gibt es eine Möglichkeit Kodi wieder als Player, aber extern einzusetzen? (mit extern meine ich ein separater Rechner wie NUC o.ä.)
Hier wird eben gefragt dass das Signal unverändert an madVR geschickt wird.
Mein Z9X ist zwar OK zum schauen, macht seine Arbeit sauber, die Bildqualität ist auch gut, aber die Menus sind ein Graus. Richtig langsam.
Die Frage verstehe ich nicht ganz.
Du lässt das processing demnächst madvr machen, soweit ok.
Von was kommst du und wieso möchtest du madvr vom Player getrennt laufen lassen ?
KODI ist ja eigentlich ein Mediencenter/Server, ja da ist auch ein Player integriert, aber der ist nicht das gelbe vom Ei und nicht mit madvr
kompatibel, sprich wenn KODI+madvr, dann musst du auch eine anderen Player einsetzen MPC, MPV o.ä.
-
Vielleicht lesen hier mehr Nutzer mit, deswegen stelle ich meine Frage hier nochmal:
ZitatDie Leinwand hört sich auch für mich sehr interessant an, kann mir jemand etwas zu den Tüchern sagen ?
Die meisten nehmen ja die AT Version mit Sound Max4K Tuch, da ich keine Verwendung für ein AT Tuch habe,
würde mir das WG1PRO Tuch reichen.
Hat das zufällig jemand im Einsatz ?
Aktuell nutze ich eine uralte 16:9 Rollo Leinwand mit Gain 1.3 im 100" Format.
Quelle: Link
-
Ich glaub da das Board ja nur den vorletzten Sockel 1151 unterstützt, wäre imho ein späteres Komplett-Upgrade,
sinnvoller.
Ist das Bild "so schlecht", dass es eine stärkere GPU unbedingt benötigt ?
Evtl. würde ich dann auf dem Gebrauchtmarkt nachsehen, wenn die 40er jetzt gut verfügbar sind,
dürften ältere 30er günstig abzufischen sein.
Ich überlege ja selbst meine 3060ti gegen eine 4090 zu tauschen (man gönnt sich ja sonst nix ),
aber dann gehts wieder los mit neuem NT wegen den HPW Anschlüssen usw.
Im SVP Thread hat auch ein Nutzer gesagt das in Zusammenspiel mit SVP der Speicher schnell die Limitierung
wird, dort wurde dann direkt DDR5 als sinnvoll erachtet.
Wenn man sich dann auf DDR6 RAM vorbereitet (geplant 2024)....