mrnorman / miniweatherml Goto Github PK
View Code? Open in Web Editor NEWExploring Machine Learning methods and workflows in a simplified weather model
License: BSD 2-Clause "Simplified" License
miniweatherml's Issues
hdf5 flags for Summit config file
The MW_LINK_FLAGS in the Summit config file may need to be updated as
export MW_LINK_FLAGS="`${OLCF_NETCDF_C_ROOT}/bin/nc-config --libs` -L/ccs/home/imn/yaml/lib64 -lyaml-cpp -L${OLCF_PARALLEL_NETCDF_ROOT}/lib -lpnetcdf -lyaml-cpp -lhdf5_hl -lhdf5"
"inference_ponni" Supercell Kesler Not Working
Hello,
I am exploring the supercell_kessler_surrogate example, and have successfully built it, but the actual simluation isn't working correctly.
I made one change to the microphysics_physics_ponni.h file to print more information about the domain:
diff --git a/experiments/supercell_kessler_surrogate/custom_modules/microphysics_kessler_ponni.h b/experiments/supercell_kessler_surrogate/custom_modules/microphysics_kessler_ponni.h
index 4b4c862..571f206 100644
--- a/experiments/supercell_kessler_surrogate/custom_modules/microphysics_kessler_ponni.h
+++ b/experiments/supercell_kessler_surrogate/custom_modules/microphysics_kessler_ponni.h
@@ -201,10 +201,6 @@ namespace custom_modules {
rho_r_tmp(k,i) = std::max( 0._fp , ponni_out( 3 , iglob ) * (scl_out(3,1) - scl_out(3,0)) + scl_out(3,0) );
});
- // std::cout << yakl::intrinsics::maxval( temp_tmp ) << "\n:";
- // std::cout << yakl::intrinsics::maxval( rho_v_tmp ) << "\n:";
- // std::cout << yakl::intrinsics::maxval( rho_c_tmp ) << "\n:";
- // std::cout << yakl::intrinsics::maxval( rho_r_tmp ) << "\n:";
/////////////////////////////////////////////////////////////////////////
// NORMAL
@@ -263,6 +259,14 @@ namespace custom_modules {
temp_diff (k,i) = temp_tmp (k,i) - temp (k,i);
});
+ std::cout << "Max rho_v " << yakl::intrinsics::maxval(rho_v) << "\n";
+ std::cout << "Max rho_c " << yakl::intrinsics::maxval(rho_c) << "\n";
+ std::cout << "Mac rho_r " << yakl::intrinsics::maxval(rho_r) << "\n";
+
+ std::cout << "Sum rho_v " << yakl::intrinsics::sum(rho_v) << "\n";
+ std::cout << "Sum rho_c " << yakl::intrinsics::sum(rho_c) << "\n";
+ std::cout << "Sum rho_r " << yakl::intrinsics::sum(rho_r) << "\n";
+
std::cout << "Relative diff rho_v: " << yakl::intrinsics::sum( rho_v_diff ) / rho_v_diff.size() << "\n";
std::cout << "Relative diff rho_c: " << yakl::intrinsics::sum( rho_c_diff ) / rho_c_diff.size() << "\n";
std::cout << "Relative diff rho_r: " << yakl::intrinsics::sum( rho_r_diff ) / rho_r_diff.size() << "\n";When I run: mpirun -np 1 ./inference_ponni inputs/input_euler3d.yaml
it produces the output:
Using YAKL Timers
Using memory pool. Initial size: 4.92383GB ; Grow size: 4.92383GB.
NVIDIA A100-SXM4-40GB
Inference model has 5 layers -- with 104 total trainable parameters.
1: Matvec with 5 inputs, 10 outputs, and 50 trainable parameters
2: Bias with 10 inputs, 10 outputs, and 10 trainable parameters
3: Relu with 10 inputs, 10 outputs, and 0 trainable parameters
4: Matvec with 10 inputs, 4 outputs, and 40 trainable parameters
5: Bias with 4 inputs, 4 outputs, and 4 trainable parameters
For Array labeled: "Unlabeled: YAKL_DEBUG CPP macro not defined"
Number of Dimensions: 2
Total Number of Elements: 10
Dimension Sizes: 5, 2,
201.819 302.293 0.0929455 1.14418 0 0.0194617 0 0.00439983 0 0.015579
For Array labeled: "Unlabeled: YAKL_DEBUG CPP macro not defined"
Number of Dimensions: 2
Total Number of Elements: 8
Dimension Sizes: 4, 2,
201.819 302.293 0 0.0194617 0 0.00441832 0 0.0155779
Max rho_v 0.0155121
Max rho_c 0.000127301
Mac rho_r -0
Sum rho_v -nan
Sum rho_c 0.113297
Sum rho_r -nan
Relative diff rho_v: -nan
Relative diff rho_c: -9.06249e-06
Relative diff rho_r: -nan
Relative diff temp : -nan
Max rho_v -1.79769e+308
Max rho_c -1.79769e+308
Mac rho_r -1.79769e+308
Sum rho_v -nan
Sum rho_c -nan
Sum rho_r -nan
Relative diff rho_v: -nan
Relative diff rho_c: -nan
Relative diff rho_r: -nan
...It seems like the domain is not initialized corretly to begin with? I'm not sure what's going on, or if I have done something wrong. I'm happy to provide any additional information you may need.
Change is_active in kessler surrogate to have `||` instead of `&&`
Add coupler snapshots + dataGen modules
Coupler snapshots will take in a std:vectorstd::string varnames. Then create a DataManager with just those and pass it back.
dataGen modules will take (coupler,dt_phys,snapshot1,snapshot2)
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.