issues
search
graphcore-research
/
pyscf-ipu
PySCF on IPU
https://github.com/graphcore-research/pyscf-ipu#readme
Apache License 2.0
42
stars
2
forks
source link
Similarities between DFT and Deep Learning
#85
Open
AlexanderMath
opened
1 year ago
AlexanderMath
commented
1 year ago
List of similarities between DFT and Deep Learning
Both solve optimization problem
Both require initialization techniques (e.g. minao init or xavier init).
Both utilize momentum techniques (e.g. DIIS or Adam)
The tensor that goes through the computational graph gets transformed
x = x + f(x)
(e.g. density mixing or residual connections)
Both use mixed precision (e.g. terachem keeps large ERIs in f64 and the remainder in f32, torch has automatic mixed precision with torch.amp)
Both uses gradients (e.g. manually implemented derivatives for forces or automatically with torch.grad)
List of similarities between DFT and Deep Learning
x = x + f(x)
(e.g. density mixing or residual connections)