MCPcopy Create free account
hub / github.com/MotrixLab/MotionDiffuse / LossType

Class LossType

text2motion/models/gaussian_diffusion.py:300–309  ·  view source on GitHub ↗

Source from the content-addressed store, hash-verified

298
299
300class LossType(enum.Enum):
301 MSE = enum.auto() # use raw MSE loss (and KL when learning variances)
302 RESCALED_MSE = (
303 enum.auto()
304 ) # use raw MSE loss (with RESCALED_KL when learning variances)
305 KL = enum.auto() # use the variational lower-bound
306 RESCALED_KL = enum.auto() # like KL, but rescale to estimate the full VLB
307
308 def is_vb(self):
309 return self == LossType.KL or self == LossType.RESCALED_KL
310
311
312class GaussianDiffusion:

Callers

nothing calls this directly

Calls

no outgoing calls

Tested by

no test coverage detected