TorchVinecop.cdf
- TorchVinecop.cdf(u, *, N=10000, qrng=True, num_threads=1, seeds=None, x=None, block_size=4096)
Evaluate the joint CDF at each query row via quasi-Monte-Carlo.
- Parameters:
- uarray, shape (m, d), dtype float
Query points in
[0, 1]^d.- Nint, default=10000
Number of Monte-Carlo samples.
- qrngbool, default=True
Draw quasi-random samples (matches
cdf()).- num_threadsint, default=1
Accepted for parity; ignored.
- seedslist of int or None, optional
RNG seeds forwarded to
simulate().- xarray or None, optional
Must be
None— a per-row conditional CDF is not supported (the Monte-Carlo dominance estimate cannot condition each query row on a different covariate without per-xresampling).- block_sizeint, default=4096
Query rows processed per iteration (peak-memory control).
- Returns:
- array, shape (m,), dtype float
CDF values in
[0, 1].
- Raises:
- NotImplementedError
If
xis notNone.