globus-labs cascade issues

globus-labs / cascade

Fast and reliable dynamic simulations by concurrently learning, auditing, and using surrogates

MIT License

1 stars 0 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Add an interface to schnetpack

#60 WardLT opened 2 weeks ago
0
Very different results moving to mace-torch==0.3.7

#59 WardLT opened 3 weeks ago
1
Minor changes to the training script

#58 WardLT closed 2 weeks ago
1
WIP: use mace as a target

#57 miketynes closed 3 weeks ago
0
Bug fix: MACE required results for inference

#56 WardLT closed 2 months ago
1
Set up runs on perlmutter

#55 miketynes opened 2 months ago
0
Introduce a MACE model interface

#54 WardLT closed 2 months ago
1
Add option to begin training from initial models

#53 WardLT closed 2 months ago
1
Handle blending of stress calculations

#52 miketynes closed 3 months ago
1
Better handle uncertainty with pretrained models

#51 WardLT opened 3 months ago
0
Provide more options for how to update model during on-the-fly learning

#50 WardLT closed 2 months ago
0
Add CHGNet

#49 WardLT closed 3 months ago
1
Add LDA calculator template

#48 WardLT closed 3 months ago
1
first pass implement ML/target blending

#47 miketynes closed 3 months ago
2
Set up computations for quality of MD runs

#46 WardLT opened 4 months ago
2
Make the ensemble a choice

#45 WardLT closed 4 months ago
0
Some Dynamics Equations Are a Problem

#44 WardLT opened 4 months ago
0
Introduce changes for problems in first round of on-the-fly training

#43 WardLT closed 4 months ago
1
Energy offset at the end of TorchANI training

#42 WardLT opened 4 months ago
0
Change to different NPT, record start time/git hash

#41 WardLT closed 4 months ago
0
Record git commit before run

#40 WardLT closed 4 months ago
0
Switch to different NPT for Proxima runs

#39 WardLT closed 4 months ago
0
Pretrained models, better tracking for Proxima

#38 WardLT closed 4 months ago
1
Introduce ability to reduce proxima training size

#37 WardLT closed 4 months ago
1
Improve ability to tune hyperparameters of ANI model

#36 WardLT closed 4 months ago
1
Track the model version number, invocation distribution in Proxima

#35 WardLT closed 4 months ago
0
Fix ASE calculator for TorchANI

#34 WardLT closed 4 months ago
1
Problem with stresses for the Proxima calculator

#33 WardLT closed 4 months ago
0
Adjust for the number of atoms in batch correctly

#32 WardLT closed 4 months ago
1
Start from a pre-trained model for Proxima run

#31 WardLT closed 4 months ago
0
Restore code for looking up template

#30 WardLT closed 4 months ago
1
Reduce training set sizes during Proxima run

#29 WardLT closed 4 months ago
0
Add stress to the loss function

#28 WardLT closed 4 months ago
1
Bug fixes for Proxima implementation

#27 WardLT closed 4 months ago
1
Add preconditioner for TorchANI

#26 WardLT closed 4 months ago
1
Add Scaling Layer to TorchANI Models

#25 WardLT closed 4 months ago
0
Add metrics tracking to TorchANI

#24 miketynes closed 4 months ago
1
Allow EnsembleCalculator to store more properties

#23 WardLT closed 5 months ago
2
Implement a serial version of learning on the fly

#22 WardLT closed 4 months ago
7
Case: surrogate gets stuck

#21 miketynes opened 5 months ago
0
Introduce an interface for running dynamics

#20 WardLT closed 5 months ago
9
EnsembleDriftAuditor

#19 miketynes opened 5 months ago
1
Method for Updating threshold in auditor

#18 miketynes opened 5 months ago
0
Prototype ensemble and auditor classes

#17 miketynes closed 5 months ago
0
Start a module for training surrogates

#16 WardLT closed 5 months ago
0
Initial "juggling" application

#15 WardLT opened 5 months ago
0
Write an interface for "should we trust this segment"

#14 WardLT opened 5 months ago
0
Add a faster machine learning surrogate

#13 WardLT closed 5 months ago
0
Remove ase timeout

#12 miketynes closed 4 months ago
1
add computation of reference forces

#11 miketynes closed 5 months ago
0