Physics Informed Point Net #875
Conversation
| @@ -100,4 +100,10 @@ end | |||
| include("dgm_test.jl") | |||
| end | |||
| end | |||
|
|
|||
| if GROUP == "All" || GROUP == "PIPN" | |||
There was a problem hiding this comment.
The .github/workflows/ci.yml needs to add a PIPN group as well for this to run.
|
Looks almost good to go! Left a few comments. |
|
Rebase |
|
Thanks! I will take a look as well in a couple of days. |
Co-authored-by: Christopher Rackauckas <[email protected]>
ChrisRackauckas
force-pushed
the
issue-#495
branch
from
68424b7
to
d4498fb
Compare
| @@ -1,7 +1,7 @@ | |||
| """ | |||
| $(DocStringExtensions.README) | |||
| """ | |||
| module NeuralPDE | |||
| module NeuralPDE | |||
There was a problem hiding this comment.
| module NeuralPDE | |
| module NeuralPDE |
| @@ -55,7 +55,7 @@ include("BPINN_ode.jl") | |||
| include("PDE_BPINN.jl") | |||
| include("dgm.jl") | |||
|
|
|||
| export NNODE, NNDAE, | |||
| export NNODE, PIPN, init_pipn_params, NNDAE, | |||
There was a problem hiding this comment.
| export NNODE, PIPN, init_pipn_params, NNDAE, | |
| export NNODE, PIPN, init_pipn_params, NNDAE, |
| @@ -1,7 +1,7 @@ | |||
| """ | |||
| ??? | |||
| """ | |||
| struct LogOptions | |||
| struct LogOptions | |||
There was a problem hiding this comment.
| struct LogOptions | |
| struct LogOptions |
|
|
||
|
|
|
|
||
| ## Positional Arguments | ||
|
|
||
| * `chain`: a Lux chain specifying the overall network architecture. The input and output dimensions |
There was a problem hiding this comment.
Is this even needed as shared mlps are constructed inside?
| global_feature_repeated = repeat(global_feature, 1, size(point_features, 2)) | ||
| combined_features = vcat(point_features, global_feature_repeated) | ||
| combined_features, st3 = Lux.apply(model.after_pool_mlp, combined_features, ps.after_pool_mlp, st.after_pool_mlp) | ||
| output, st4 = Lux.apply(model.final_layer, combined_features, ps.final_layer, st.final_layer) |
There was a problem hiding this comment.
Doesn't the original architecture also have another block of shared mlps?
| flat_ps = ComponentArray(ps) | ||
| new_flat_ps = typeof(flat_ps)(θ) | ||
| return ComponentArray(new_flat_ps) | ||
| end |
There was a problem hiding this comment.
| end | |
| end | |
| @@ -1,4 +1,4 @@ | |||
| using Pkg | |||
| using Pkg | |||
There was a problem hiding this comment.
| using Pkg | |
| using Pkg |
|
|
||
| @test prob isa OptimizationProblem | ||
| end | ||
| end |
There was a problem hiding this comment.
| end | |
| end | |
| strategy = GridTraining(0.1) | ||
| discretization = PhysicsInformedNN(chain, strategy) | ||
|
|
||
| prob = discretize(pde_system, discretization) |
There was a problem hiding this comment.
Is it working with solve?
There was a problem hiding this comment.
Ok, cool. Can you add it to the test and also address other comments?
Checklist
contributor guidelines, in particular the SciML Style Guide and
COLPRAC.
Additional context
Added PIPNs to NeuralPDE
Fixes: #495