Compute an unbiased estimate to the log-likelihood in bits/dim. Args: cond: dict with key state/rgb; more recent obs at the end state: (B, To, Do) data: (B x Ta x Da) Returns: logprob: (B,)
(
model,
model_ft,
data,
denoising_steps,
ft_denoising_steps,
cond,
**kwargs,
)
source not stored for this graph (policy: none)
nothing calls this directly
no test coverage detected