|
|
03:41:12 PM INFO startup: first paint 8600 ms (pre-window 6979 ms, ui build 1609 ms)
03:41:28 PM INFO Processing started by user
03:41:28 PM INFO Output folder: F:/Downloads
03:41:28 PM INFO Output pattern: {original}_restored.mp4
03:41:28 PM INFO Files queued: 1
03:41:29 PM INFO Processing started
03:41:29 PM INFO Started processing MIDA-849-1min-TEST.mp4
03:41:31 PM INFO Loads checkpoint by local backend from path: E:\VideoEdition\jasna-windows-0.11.3\model_weights\lada_mosaic_restoration_model_generic_v1.2.pth
03:42:33 PM INFO BasicVSR++ using TRT sub-engines (fp16=True)
03:42:33 PM INFO RestorationPipeline: secondary=none denoise=NONE denoise_step=AFTER_PRIMARY
03:42:33 PM INFO Restoration models loaded (reused across video jobs)
03:42:38 PM INFO RF-DETR detection model loaded: E:\VideoEdition\jasna-windows-0.11.3\model_weights\rfdetr-v6.bs1-4.fp16.win.engine (batch_size=4)
03:42:51 PM INFO VR mode: requested=auto resolved=off projection=none reason=display aspect 1.778 is outside the SBS gate
03:42:51 PM INFO VramOffloader: threshold=15552 MiB (total=16302 MiB, safetynet=750 MiB)
03:42:56 PM INFO Using VALI NVDEC decoder for F:\Downloads\MIDA-849-1min-TEST.mp4
03:42:56 PM INFO Processing F:\Downloads\MIDA-849-1min-TEST.mp4: 1850 frames @ 29.97002997002997 fps, 1920x1080
03:42:56 PM INFO Using VALI NVDEC decoder for F:\Downloads\MIDA-849-1min-TEST.mp4
03:43:25 PM INFO [timing] decode-detect: decode=2.3s, detect-track=26.5s (tracked 28.8s)
03:43:25 PM INFO [timing] primary: queue-wait=12.5s, restore=20.2s, queue-put=0.0s (tracked 32.8s)
03:43:25 PM INFO [timing] secondary: queue-wait=32.1s, restore=0.1s, queue-put=0.6s (tracked 32.8s)
03:43:25 PM INFO [timing] blend-encode: queue-wait=6.6s, decode=4.4s, result-wait=7.1s, blend=10.0s, write=1.3s (tracked 29.3s)
03:43:26 PM INFO VRAM — min: 2933 MiB, max: 4484 MiB, avg: 3771 MiB | offloads: 0, total offloaded: 0 MiB
03:43:26 PM INFO VRAM usage at end — 3444.2 MiB
03:43:26 PM INFO RAM usage at end — 3872.1 MiB
03:43:26 PM INFO Finished processing MIDA-849-1min-TEST.mp4
03:43:27 PM INFO Restoration models unloaded
03:43:27 PM INFO Processing completed
03:43:27 PM INFO All jobs completed
03:44:26 PM INFO Processing started by user
03:44:26 PM INFO Output folder: F:/Downloads
03:44:26 PM INFO Output pattern: {original}_restored.mp4
03:44:26 PM INFO Files queued: 1
03:44:26 PM INFO Processing started
03:44:26 PM INFO Started processing MIDA-849-1min-TEST.mp4
03:44:27 PM INFO Restoration models loaded (reused across video jobs)
03:44:27 PM INFO RF-DETR detection model loaded: E:\VideoEdition\jasna-windows-0.11.3\model_weights\rfdetr-v6.bs1-4.fp16.win.engine (batch_size=4)
03:44:28 PM INFO VR mode: requested=auto resolved=off projection=none reason=display aspect 1.778 is outside the SBS gate
03:44:28 PM INFO LTX speed test: no speed profile for this GPU and model yet
03:44:28 PM INFO Using VALI NVDEC decoder for F:\Downloads\MIDA-849-1min-TEST.mp4
03:44:31 PM INFO LTX plan: 209 frames, 0 scene cuts, 2 tracks, 4 windows
03:49:46 PM INFO Using VALI NVDEC decoder for F:\Downloads\MIDA-849-1min-TEST.mp4
03:50:38 PM ERROR Failed to process MIDA-849-1min-TEST.mp4: LTX inference failed in the protected worker: InductorError: UnicodeDecodeError: 'gbk' codec can't decode byte 0x87 in position 1562: illegal multibyte sequence
Set TORCHDYNAMO_VERBOSE=1 for the internal stack trace (please do this especially if you're reporting a bug to PyTorch). For even more developer context, set TORCH_LOGS="+dynamo"
(output_code.py:545 <- compile_fx.py:1753 <- compile_fx.py:1836 <- video_encoder.py:158)
Traceback (most recent call last):
File "E:\VideoEdition\jasna-windows-0.11.3\jasna\gui\processor.py", line 256, in _process_job
File "E:\VideoEdition\jasna-windows-0.11.3\jasna\gui\processor.py", line 342, in _run_pipeline
File "E:\VideoEdition\jasna-windows-0.11.3\jasna\gui\processor.py", line 435, in _run_video_job
File "E:\VideoEdition\jasna-windows-0.11.3\jasna\pipeline.py", line 605, in run
File "E:\VideoEdition\jasna-windows-0.11.3\jasna\pipeline.py", line 307, in _run_ltx
File "E:\VideoEdition\jasna-windows-0.11.3\jasna\ltx\restore.py", line 1133, in restore_video
File "E:\VideoEdition\jasna-windows-0.11.3\jasna\ltx\restore.py", line 1067, in restore_spans
File "E:\VideoEdition\jasna-windows-0.11.3\jasna\ltx\restore.py", line 1087, in _restore_spans
File "E:\VideoEdition\jasna-windows-0.11.3\jasna\ltx\restore.py", line 890, in prepare_spans
File "E:\VideoEdition\jasna-windows-0.11.3\jasna\ltx\restore.py", line 607, in encode_span
File "E:\VideoEdition\jasna-windows-0.11.3\jasna\ltx\restore.py", line 576, in frame
File "E:\VideoEdition\jasna-windows-0.11.3\jasna\ltx\inference.py", line 43, in __call__
File "E:\VideoEdition\jasna-windows-0.11.3\jasna\ltx\inference.py", line 106, in request
File "E:\VideoEdition\jasna-windows-0.11.3\jasna\protection\ltx_worker.py", line 304, in request
File "E:\VideoEdition\jasna-windows-0.11.3\jasna\protection\ltx_worker.py", line 254, in _exchange
jasna.protection.ltx_worker.WorkerError: LTX inference failed in the protected worker: InductorError: UnicodeDecodeError: 'gbk' codec can't decode byte 0x87 in position 1562: illegal multibyte sequence
Set TORCHDYNAMO_VERBOSE=1 for the internal stack trace (please do this especially if you're reporting a bug to PyTorch). For even more developer context, set TORCH_LOGS="+dynamo"
(output_code.py:545 <- compile_fx.py:1753 <- compile_fx.py:1836 <- video_encoder.py:158)
03:50:38 PM INFO Restoration models unloaded
03:50:38 PM INFO Processing completed
03:50:38 PM INFO All jobs completed
|
|
发表于 
楼主