Rife-App 1.4.3: A few important fixes

I'm trying to improve this on next updates, but you can try setting batch =1 to save some ram for now.

GRisk

For a lossless output, will add later on the option to export as images, similar to Dain-App. This crash seen like a bug, will try to fix it along other stuff on the next update. X3 X5 X6 should be possible, but it would take the same time to render as the next option; X3 same time as X4, X5 same time as X8, X6 same time as X8, so not sure how usefull would be, plus the result would be very similar.

GRisk

there is a button that says download

QLST

hi there, how do I download it? I wanted to try it on my video

It gets to 55GB of RAM usage and then gives the error

Tried with 1.3 - and still got this: Starting... ['F:/RIFE-App 1.4.3/R_.kava.mov'] Input FPS: 25.0 F:/RIFE-App 1.4.3 Main Folder: F:/RIFE-App 1.4.3/R_/ Starting PNG frames extraction! Finished PNG frames extraction! Using Benchmark: True Using Half-Precision: True Batch Size: -1 Selected auto batch size, testing a good batch size. Setting new batch size to 2 Starting Interpolation! 38%|██████████▌ | 82/217 [03:29<05:27, 2.42s/it, file=File 39]Traceback (most recent call last): File "my_design.py", line 72, in run File "my_DAIN_class.py", line 1298, in RenderVideo File "my_DAIN_class.py", line 1095, in StepRenderInterpolation File "torch\utils\data\dataloader.py", line 435, in __next__ data = self._next_data() File "torch\utils\data\dataloader.py", line 1085, in _next_data return self._process_data(data) File "torch\utils\data\dataloader.py", line 1111, in _process_data data.reraise() File "torch\_utils.py", line 428, in reraise raise self.exc_type(msg) RuntimeError: Caught RuntimeError in DataLoader worker process 0. Original Traceback (most recent call last): File "torch\utils\data\_utils\worker.py", line 198, in _worker_loop data = fetcher.fetch(index) File "torch\utils\data\_utils\fetch.py", line 44, in fetch data = [self.dataset[idx] for idx in possibly_batched_index] File "torch\utils\data\_utils\fetch.py", line 44, in data = [self.dataset[idx] for idx in possibly_batched_index] File "DainDataset.py", line 64, in __getitem__ RuntimeError: [enforce fail at ..\c10\core\CPUAllocator.cpp:73] data. DefaultCPUAllocator: not enough memory: you tried to allocate 353894400 bytes. Buy new RAM! Exception ignored in thread started by: Traceback (most recent call last): File "my_DAIN_class.py", line 102, in queue_file_save RuntimeError: [enforce fail at ..\c10\core\CPUAllocator.cpp:73] data. DefaultCPUAllocator: not enough memory: you tried to allocate 353894400 bytes. Buy new RAM!

Gotta say I'm very impressed with Rife-App current progress, coming from Dain-App). Still works nicely with Win7+1080ti combo - much faster with less crashes. "Concat original video at the side of output" is a nice feature as well. Overall quality of output seem to improved a bit, doesn't seem completely "lossless" compared to input though. Although I haven't seen similar artifacts like before which is nice. Changing "frame similarity" percent value usually ends up with a crash (something like Error 23, didn't look in it that much). Do you have plans on adding "weird" interpolation values like X3,X5,X6? Or it's really not necessary with current implementation of Animation mode/will be hard to implement (worse output in general)?

Sergey Tokarev

It seen that your graphic card is having problems allocating memory to handle the frames, what card you have? Can you try to set the batch size to 1 and see if it work? Edit: Just seen that you already using batch = 1. Can you try it a lower resolution video?

GRisk

Using Benchmark: True Using Half-Precision: True Batch Size: 1 Input FPS: 59.94006309148265 Starting Interpolation! 1%|▎ | 268/20241 [12:01<21:56:48, 3.96s/it, file=File 266]THCudaCheck FAIL file=..\aten\src\THC\THCCachingHostAllocator.cpp line=278 error=2 : out of memory Traceback (most recent call last): File "my_design.py", line 72, in run File "my_DAIN_class.py", line 1561, in RenderVideo File "my_DAIN_class.py", line 1312, in StepRenderInterpolation File "torch\utils\data\dataloader.py", line 435, in __next__ data = self._next_data() File "torch\utils\data\dataloader.py", line 1085, in _next_data return self._process_data(data) File "torch\utils\data\dataloader.py", line 1111, in _process_data data.reraise() File "torch\_utils.py", line 428, in reraise raise self.exc_type(msg) RuntimeError: Caught RuntimeError in pin memory thread for device 0. Original Traceback (most recent call last): File "torch\utils\data\_utils\pin_memory.py", line 31, in _pin_memory_loop data = pin_memory(data) File "torch\utils\data\_utils\pin_memory.py", line 55, in pin_memory return [pin_memory(sample) for sample in data] File "torch\utils\data\_utils\pin_memory.py", line 55, in return [pin_memory(sample) for sample in data] File "torch\utils\data\_utils\pin_memory.py", line 47, in pin_memory return data.pin_memory() RuntimeError: cuda runtime error (2) : out of memory at ..\aten\src\THC\THCCachingHostAllocator.cpp:278

QLST

Love the work, but it just got stuck as starting interpolation and there is no video in the output folder at all. Keep up the amazing work

QLST