snrv.DatasetSnrv

class snrv.DatasetSnrv(data, lag, ln_pathweight)

Custom dataset for Snrv class

Parameters

• data (float tensor (single traj) or list of float tensors (multi traj); dim 0 = steps, dim 1 = features) – time-continuous trajectories

• lag (int) – lag in steps to apply to data trajectory

• ln_pathweight (torch.tensor, n, n = observations) –

  accumulated sum of the log Girsanov path weights between frames in the trajectory;

  Girsanov theorem measure of the probability of the observed sample path under a target potential relative to that which was actually observed under the simulation potential; identically unity (no reweighting rqd) for target potential == simulation potential and code as None

Variables

• self.lag (int) – lag in steps

• self.x_t0 (float tensor, n x dim, n = observations, dim = dimensionality of trajectory featurization) – non-time-lagged trajectory

• self.x_tt (float tensor, n x dim, n = observations, dim = dimensionality of trajectory featurization) – time-lagged trajectory

• self.pathweight (float tensor, n = observations) – pathweights from Girsanov theorem between time lagged observations; identically unity (no reweighting rqd) for target potential == simulation potential; if ln_pathweight == None => pathweight == ones

__init__(data, lag, ln_pathweight)

Methods

__init__(data, lag, ln_pathweight)