Closed rhystz closed 1 month ago
I also couldn't get the down(load) and load cog model node to work so just git cloned from thud into models/cogvideo
The result is good on 49 frames, and i don't get the "[VideoHelperSuite] - WARNING - Output images were not of valid resolution and have had padding applied" msg.
I also get grid results,
I think the I2V model simply only works with 49 frames, those are the results with other frame amounts I've tried as well.
Ahh ok great, thanks.
Hi Kijai, Thanks heaps for everything you do. I'm getting a gridded output (16 frames, otherwise same as example workflow settings) and a few inconsistently appearing msgs:
torch.Size([1, 4, 16, 60, 90]) tensor(0.) tensor(1.) [VideoHelperSuite] - WARNING - Output images were not of valid resolution and have had padding applied Prompt executed in 549.20 seconds [Crystools ERROR] For some reason, pynvml is not working in a laptop with only battery, try to connect and turn on the monitor not always there [Crystools ERROR] Monitor of GPU is turning off (not on UI!) not always there got prompt latents torch.Size([1, 8, 16, 60, 90]) Temporal tiling disabled latents.shape torch.Size([1, 8, 16, 60, 90])
I've hit OOMs a couple of times but seem to be able to move forward straight afterwards. im on a laptop rtx a4500. Do you know what could be causing it? Non-I2V 5b model runs well.