A CUDA kernel extension for efficient Inside Algorithm built on the genbmm library. It computes the inside score on the diagonals iteratively.
pip install git+https://github.com/lyutyuh/genbmm
genbmm.logbmminside_rule(inside, rule, width)
computes the following values for all
genbmm.logbmminside(inside, width)
computes the following values for all
This kernel is used in structured span selector.
https://github.com/lyutyuh/structured-span-selector/blob/main/outside_mp.py#L62
This library is a collection of missing matrix-multiply like operations for PyTorch. It was developed to provide operators needed for PyTorch-Struct.
The library has currently has two components. It only supports CUDA operations.
- Generalized matrix-multiplication with gradients (log-space, max, sample)
- Banded sparse matrices
pip install git+https://github.com/lyutyuh/genbmm
Computing matrix multiplies over non-standard semi-rings in PyTorch requires creating large intermediary terms with views. This is particularly bad because they are stored internally for backprop. This library implements some alternative matrix multiplies in CUDA to avoid this issue.
import genbmm
a = torch.rand(10, 3, 4).cuda().requires_grad_(True)
b = torch.rand(10, 4, 5).cuda().requires_grad_(True)
# Log-Sum-Exp
c = genbmm.logbmm(a, b)
# Equivalent
a = a.unsqueeze(-1)
b = b.unsqueeze(-3)
c2 = (a + b).logsumexp(-2)
# Grad
prob_a, prob_b = torch.autograd.grad(c.sum(), (a, b))
# Max
c = genbmm.maxbmm(a, b)
# Equivalent
a = a.unsqueeze(-1)
b = b.unsqueeze(-3)
c2, = (a + b).max(-2)
# Grad
argmax_a, argmax_b = torch.autograd.grad(c.sum(), (a, b))
# Sample
c = genbmm.samplebmm(a, b)
# Equivalent
a = a.unsqueeze(-1)
b = b.unsqueeze(-3)
c2 = (a + b).logsumexp(-2)
# Grad
sample_a, sample_b = torch.autograd.grad(c.sum(), (a, b))
# c2 = (a + b).softmax(-2).sample(-2)
# Product-Max
c = genbmm.prodmaxbmm(a, b)
# Equivalent
a = a.unsqueeze(-1)
b = b.unsqueeze(-3)
c2, = (a * b).max(-2)
# Grad
grad_a, grad_b = torch.autograd.grad(c.sum(), (a, b))
See https://nbviewer.jupyter.org/github/harvardnlp/genbmm/blob/master/notebooks/Sparse.ipynb.