[ADD] PyTorch embedding layer to reduce memory footprint - #451

Open

ravinkohli wants to merge 22 commits into

reg_cocktailsfrom

reg_cocktails-pytorch_embedding

Commits on Oct 25, 2022

update eta for experiments
ravinkohli
committed
0bd7d5f
add check if True is in value range
ravinkohli
committed
76dae54
Reg cocktails common paper modifications 2 (#417)
ravinkohli
committed
d26c611
have working embedding from pytroch
ravinkohli
committed
bc60e31
divide columns to encode and embed based on threshold
ravinkohli
committed
13fad76
cleanup unwanted changes
ravinkohli
committed
cf4fd98
use shape after preprocessing in base network backbone
ravinkohli
committed
af41dd7
remove redundant call to load datamanager
ravinkohli
committed
9706875
add init file for column splitting
ravinkohli
committed
def144c
fix tests
ravinkohli
committed
926a757
fix precommit and add test changes
ravinkohli
committed
7567d26
[ADD] Calculate memory of dataset after one hot encoding (pytorch embedding) (#437)

ravinkohli
and
nabenabe0928
committed
09fdc0d
suggestions from review
ravinkohli
committed
3aef02e
add preprocessed_dtype to determine double or float
ravinkohli
committed
8e3dbef
test fix in progress
ravinkohli
committed
52427bc
TODO: fix errors after rebase
ravinkohli
committed
90512ee
Reg cocktails apt1.0+reg cocktails pytorch embedding reduced (#454)
ravinkohli
committed
895b904
fix embeddings after rebase
ravinkohli
committed
033bca7
fix error with pytorch embeddings
ravinkohli
committed
d4cd8b4
fix redundant code
ravinkohli
committed
a5807cb
change userdefined to False
ravinkohli
committed
960e1ef
remove using categorical columns
ravinkohli
committed
1be80d5