Closed janfb closed 1 month ago
I would also be interested in doing this!
me too!
I am not part of the org yet, though would like to be assigned.
rough work plan for the "basic" components:
neural_nets
, something like a VectorFieldEstimator
super class that is the generic class for specific subclasses, including ScoreEstimator
and FlowMatchEstimator
, corresponding to score matching (i.e., denoising diffusion model) and flow matching, respectively, with their specific lossesVectorFieldSampler
class in samplers
that receives the VectorFieldEstimator
and samples by, e.g., solving the ODE/SDEVectorFieldPosterior
object in inference/posteriors/
that receives the VectorFieldEstimator
and sets up a VectorFieldSampler
to create the posterior distribution, similar to DirectPosterior
does this sound about right? (also @janfb, @michaeldeistler, etc.)
torchdiffeq seems like a reasonable choice for ODEs in pytorch. It's also used in dingo! Alternatives are torchDyn and torchode.
I don't know how to link the branch to the issue here but it's in score_estimator
Is your feature request related to a problem? Please describe.
Score estimation has been introduced for SBI, see
it would be great to integrate these approaches into
sbi
.Describe the solution you'd like
(feel free to move the last three points to separate issues, or create individual PRs for each).