StuvX

I am wondering about this too - I think it strips off the beginning of the "module." state_dict entries so that it matches the model state_dict. I think the issue...

You can see this in the forward pass of the HighResolutionNet module. After interpolation upsizing the resulting arrays are concatenated, and then passed through the last_layer submodule that consists of:...

I suspect this is an error with how the model is loading, you need to explicitly map the model weights to your device, otherwise whatever weights are not mapped will...

My understanding is that the speedup is largely from an optimisation of model depth balanced against image resolution - so parallel inference is run using: 1: full resolution through 4...

Ok - the channel reversal is becuase the weights are loaded from the caffe model, but i took it out and my network learned as well as if it was...

Hi Bendidi - for your script edit it seems to me that the combined training ground-truth (valid_gt) will be sent to the evaluator, and so the network will learn to...

I'm having the same problem on OSX - I think it's traced to the tensorflow-fcn module and possibly to do with the build of tensorflow you're using (I'm on r1.3)....

OK - so my particular problem is that I trained the model on one machine, and then copied it to my other machine, but when you do that the path...

Thanks - but I still receive the following error: Traceback (most recent call last): File "/taichi/massSpring.py", line 20, in mesh = Patcher.load_mesh(args.model, relations=["FV"]) File "/miniconda3/envs/taichi/lib/python3.10/site-packages/meshtaichi_patcher/__init__.py", line 32, in load_mesh return...

@BillXu2000 You can use this code to generate the PLY file: ``` import numpy as np import taichi as ti import meshtaichi_patcher as Patcher ti.init(arch=ti.cpu) coords = [ [1,0,0], [0,1,0],...