-
Notifications
You must be signed in to change notification settings - Fork 1
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
Showing
5 changed files
with
247 additions
and
20 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,184 @@ | ||
| # This file is a part of AdaptiveFlows.jl, licensed under the MIT License (MIT). | ||
|
|
||
| @recipe function plot_flow(flow::AbstractFlow, samples::AbstractMatrix; | ||
| d_sel = nothing, | ||
| n_bins_1D = 100, | ||
| n_bins_2D = 100, | ||
| colorbar = false, | ||
| h_x = true, | ||
| h_y = true, | ||
| fs = nothing, | ||
| size_plot = (1000, 1000), | ||
| p = x -> 1/sqrt(2pi) * exp(-x^2/2), | ||
| training_metadata = nothing | ||
| ) | ||
| n_dims = size(samples,1) | ||
|
|
||
| if isnothing(d_sel) | ||
| d_sel = 1:Integer(minimum([6, n_dims])) | ||
| end | ||
| n_plots = length(d_sel) | ||
|
|
||
| if isnothing(fs) | ||
| fs = n_plots > 5 ? 6 : n_plots > 3 ? 7 : 11 | ||
| end | ||
|
|
||
| samples_transformed = flow(samples) | ||
|
|
||
| stds_in = vec(std(samples, dims = 2)) | ||
| means_in = vec(mean(samples, dims = 2)) | ||
| samples_in = InvMulAdd(Diagonal(stds_in), means_in)(samples) | ||
|
|
||
| stds_out = vec(std(samples_transformed, dims = 2)) | ||
| means_out = vec(mean(samples_transformed, dims = 2)) | ||
| samples_out = InvMulAdd(Diagonal(stds_out), means_out)(samples_transformed) | ||
|
|
||
| if !isnothing(training_metadata) | ||
| training_metadata_labels = ["[Training metadata] ", "Loss: $(training_metadata[:loss]) ", "Optimizer: ", "$(training_metadata[:optimizer]) ", "# Batches: $(training_metadata[:nbatches]) ", "# Epochs: $(training_metadata[:nepochs]) ", "Sequential?: $(training_metadata[:sequential]) ", "Shuffle samples?: $(training_metadata[:shuffle_samples]) "] | ||
| end | ||
| samples_metadata = ["[Samples metadata] ", "# Samples: $(size(samples,2)) ", "# Dimensions: $n_dims ", "Displayed Dimensions: ", "$(d_sel) ", " ", " ", " "] | ||
|
|
||
| layout --> (n_plots + 1, n_plots) | ||
| size --> size_plot | ||
|
|
||
| for i in 1:n_plots | ||
| bin_range_1D = range(minimum([minimum(samples_in[i,:]), minimum(samples_out[i,:])]), stop = maximum([maximum(samples_in[i,:]), maximum(samples_out[i,:])]), length = n_bins_1D) | ||
|
|
||
| # Diagonal | ||
| subplot := i + (i-1) * n_plots | ||
| tickfontsize --> fs | ||
| labelfontsize --> fs | ||
| if h_x | ||
| @series begin | ||
| seriestype --> :stephist | ||
| bins --> bin_range_1D | ||
| normalize --> :pdf | ||
| label --> false | ||
| color --> :blue | ||
| fill --> true | ||
| alpha --> 0.3 | ||
| samples_in[d_sel[i],:] | ||
| end | ||
| end | ||
| if h_y | ||
| @series begin | ||
| seriestype --> :stephist | ||
| bins --> bin_range_1D | ||
| normalize --> :pdf | ||
| label --> false | ||
| color --> :red | ||
| fill --> true | ||
| alpha --> 0.3 | ||
| samples_out[d_sel[i],:] | ||
| end | ||
| end | ||
| @series begin | ||
| seriestype --> :line | ||
| lw --> 1.5 | ||
| color --> :black | ||
| label --> false | ||
| p | ||
| end | ||
|
|
||
| for j in i+1:n_plots | ||
| # Input, lower off-diagonal | ||
| subplot := i + (j - 1) * n_plots | ||
| @series begin | ||
| seriestype --> :histogram2d | ||
| bins --> n_bins_2D | ||
| color --> :blues | ||
| colorbar --> colorbar | ||
| background --> :white | ||
| aspect_ratio --> :equal | ||
| tickfontsize --> fs | ||
| labelfontsize --> fs | ||
| xlabel --> "x$((d_sel[i]))" | ||
| ylabel --> "x$((d_sel[j]))" | ||
| (samples_in[d_sel[i],:], samples_in[d_sel[j],:]) | ||
| end | ||
|
|
||
| # Output upper off-diagonal | ||
| subplot := j + (i - 1) * n_plots | ||
|
|
||
| @series begin | ||
| seriestype --> :histogram2d | ||
| bins --> n_bins_2D | ||
| color --> :reds | ||
| colorbar --> colorbar | ||
| aspect_ratio --> :equal | ||
| tickfontsize --> fs | ||
| labelfontsize --> fs | ||
| xlabel --> "y$(d_sel[i])" | ||
| ylabel --> "y$(d_sel[j])" | ||
| (samples_out[d_sel[i],:], samples_out[d_sel[j],:]) | ||
| end | ||
| end | ||
|
|
||
| subplot := n_plots^2 + n_plots + 1 - i | ||
|
|
||
| if i == n_plots && !isnothing(training_metadata) | ||
| for k in 1:8 | ||
| @series begin | ||
| seriestype --> :scatter | ||
| ticks --> false | ||
| framestyle --> :none | ||
| legend --> :bottomleft | ||
| legendfontsize --> fs | ||
| label --> training_metadata_labels[k] | ||
| markeralpha --> 0 | ||
| [0] | ||
| end | ||
| end | ||
| elseif i == 1 | ||
| for k in 1:8 | ||
| @series begin | ||
| seriestype --> :scatter | ||
| ticks --> false | ||
| framestyle --> :none | ||
| legend --> :bottomright | ||
| legendfontsize --> fs | ||
| label --> samples_metadata[k] | ||
| markeralpha --> 0 | ||
| [0] | ||
| end | ||
| end | ||
| else | ||
| @series begin | ||
| seriestype --> :scatter | ||
| ticks --> false | ||
| framestyle --> :none | ||
| legend --> false | ||
| markeralpha --> 0 | ||
| [0] | ||
| end | ||
| end | ||
| end | ||
| end | ||
|
|
||
| @recipe function plot_flow_res(res::Union{NamedTuple{(:result, :optimizer_state, :loss_hist, :training_metadata), Tuple{CompositeFlow, Vector{Any}, Vector{Any}, Dict{Symbol, Any}}}, | ||
| NamedTuple{(:result, :optimizer_state, :loss_hist, :training_metadata), Tuple{CompositeFlow, NamedTuple{(:flow,), Tuple{NamedTuple{(:fs,), Tuple{Vector{NamedTuple}}}}}, Vector{Float64}, Dict{Symbol, Any}}}}, | ||
| samples::AbstractMatrix; | ||
| d_sel = nothing, | ||
| n_bins_1D = 100, | ||
| n_bins_2D = 100, | ||
| colorbar = false, | ||
| h_x = true, | ||
| h_y = true, | ||
| fs = nothing, | ||
| size_plot = (1000, 1000), | ||
| p = x -> 1/sqrt(2pi) * exp(-x^2/2) | ||
| ) | ||
| @series begin | ||
| n_bins_1d := n_bins_1D | ||
| n_bins_2d := n_bins_2D | ||
| d_sel := d_sel | ||
| colorbar := colorbar | ||
| h_x := h_x | ||
| h_y := h_y | ||
| fs := fs | ||
| size := size_plot | ||
| p := p | ||
| training_metadata := res.training_metadata | ||
| (res.result, samples) | ||
| end | ||
| end |