This is the core pytorch function to call. Every torch NMR has a chainer NMR. Only fwd/bwd once per iteration.

Current: Resnet with 4 blocks (x32 spatial dim reduction) Another conv with stride 2 (x64) This is sent to 2 fc layers with final out

Outputs mesh deformations

Outputs mesh texture

Outputs [tx, ty] or [tx, ty, tz]

Virtual Trackball Control. >>> ball = Arcball() >>> ball = Arcball(initial=numpy.identity(4)) >>> ball.place([320, 320], 320) >>> bal

img, mask, kp, pose data loader

CUB Data loader

Edge length should not diverge from the original edge length. On initialization computes the current edge lengths.

Encourages minimal mean curvature shapes.

Utility to render meshes using pytorch NMR faces are F x 3 or 1 x F x 3 numpy