issues
search
cnellington
/
Contextualized
An SKLearn-style toolbox for estimating and analyzing models, distributions, and functions with context-specific parameters.
http://contextualized.ml/
GNU General Public License v3.0
65
stars
9
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Update regression.py
#48
blengerich
closed
2 years ago
0
Move models to PyTorch Lightning
#47
cnellington
closed
2 years ago
0
Sync with NOTMAD
#46
blengerich
closed
2 years ago
0
Add link_function for ContextualizedRegressor
#45
blengerich
closed
2 years ago
0
Make ContextualRegressor
#44
blengerich
closed
2 years ago
0
task-wise predict hotfix
#43
cnellington
closed
2 years ago
0
consolidate task encoder
#42
cnellington
closed
2 years ago
1
Fix predict
#41
cnellington
closed
2 years ago
0
Consolidate task encoder
#40
cnellington
closed
2 years ago
0
Flag for univariable / multivariable prediction.
#39
blengerich
closed
2 years ago
0
Plot homogeneous effects.
#38
blengerich
closed
2 years ago
1
Add plotting util for heterogeneous effects.
#37
blengerich
closed
2 years ago
1
Allow encoder to be GAM
#36
blengerich
closed
2 years ago
0
Predict returns numpy and archetype init is centered and small
#35
cnellington
closed
2 years ago
0
Allow custom activation functions
#34
blengerich
closed
2 years ago
0
Initialize with torch.rand may result in unstable training.
#33
blengerich
closed
2 years ago
0
predict_* functions should return numpy matrices, not tensors.
#32
blengerich
closed
2 years ago
0
added device specificiation for gpu/distributed training
#31
cnellington
closed
2 years ago
0
change num_archetypes default from None to 0, add paramsweep.py to si…
#30
cnellington
closed
2 years ago
0
added silent training flag
#29
cnellington
closed
2 years ago
0
Add silent training flag
#28
cnellington
closed
2 years ago
0
Regularize encoder outputs
#27
cnellington
closed
2 years ago
0
Regularize encoder outputs
#26
cnellington
closed
2 years ago
0
Correlator bootstrapping and encoder flexibility
#25
cnellington
closed
2 years ago
0
Made simulations identifiable, added true parameters for sanity checks
#24
cnellington
closed
2 years ago
0
Simulation sanity checks
#23
cnellington
closed
2 years ago
0
Update public class to new Torch module functionality
#22
cnellington
closed
2 years ago
0
added option for no archetypes in the task encoder (defaults to MLP)
#21
cnellington
closed
2 years ago
0
Added full and PCA-projected true contexts for sanity check
#20
cnellington
closed
2 years ago
0
Options for non-archetype training
#19
blengerich
closed
2 years ago
0
Saving, loading, checkpoints
#18
cnellington
closed
2 years ago
0
Added early stopping and training validation
#17
cnellington
closed
2 years ago
0
Monitor training for early stopping
#16
cnellington
closed
2 years ago
0
added native bootstrapping
#15
cnellington
closed
2 years ago
0
made task-pairs unique, tasks completely hidden from user
#14
cnellington
closed
2 years ago
0
Maybe add bootstraps natively
#13
blengerich
closed
2 years ago
0
Enable pip install, make pairwise data an internal structure
#12
cnellington
closed
2 years ago
0
Task uniqueness
#11
cnellington
closed
2 years ago
0
pip install
#10
cnellington
closed
2 years ago
0
add dynamic test loading
#9
cnellington
closed
2 years ago
0
resolves #5
#8
cnellington
closed
2 years ago
0
Loss function for Pre-training
#7
blengerich
opened
2 years ago
0
Cell-specific Contexts
#6
blengerich
closed
2 years ago
1
Dataset testsplit
#5
blengerich
closed
2 years ago
0
Distributed Training
#4
blengerich
closed
2 years ago
0
Can't visualize networks (memory)
#3
blengerich
opened
2 years ago
0
Ad hoc pairwise data loading, flexible task transformation
#2
cnellington
closed
3 years ago
0
allow X<->Y correlation (previously only X<->X)
#1
cnellington
closed
3 years ago
0
Previous