amber.plots

amber.plots.heatmap2

Reference

Original code from:

https://www.kaggle.com/threecourse/heatmap-with-change-cell-size-feature/notebook#Heatmap-with-change-cell-size-feature

Below is original header: This script is created by modifying seaborn matrix.py in https://github.com/mwaskom/seaborn, by Michael L. Waskom

heatmap2(data, vmin=None, vmax=None, cmap=None, center=None, robust=False, annot=None, fmt='.2g', annot_kws=None, cellsize=None, cellsize_vmax=None, cbar=True, cbar_kws=None, cbar_ax=None, square=False, xticklabels='auto', yticklabels='auto', mask=None, ax=None, ax_kws=None, rect_kws=None)[source]

amber.plots.ontology

utils for ontology NAS

plot_nx_dag(ss, save_fn=None)[source]

amber.plots.plot_all_stats

plot_all_stats(file_list, baseline_file)[source]

amber.plots.plotsV1

accum_opt(data, find_min)[source]
heatscatter(x, y)[source]
heatscatter_sns(x, y, figsize=(8, 8))[source]
multi_distplot_sns(data, labels, save_fn, title='title', xlab='xlab', ylab='ylab', hist=False, rug=False, xlim=None, ylim=None, legend_off=False, **kwargs)[source]
plot_action_weights(working_dir)[source]
plot_controller_hidden_states(controller, save_fn='controller_hidden_states.png')[source]
plot_controller_performance(controller_hist_file, metrics_dict, save_fn=None, N_sma=10)[source]
Example:

controller_hist_file = ‘train_history.csv’ metrics_dict = {‘acc’: 0, ‘loss’: 1, ‘knowledge’: 2}

plot_environment_entropy(entropy_record, save_fn)[source]

plot the entropy change for the state-space in the training environment. A smaller entropy indicates converaged controller.

plot_hessian(gkf, save_fn)[source]
plot_sequence_importance(model_importance, motif_importance, pos_chunks=None, neg_chunks=None, prediction=None, seq_char=None, title='sequence_importance', linespace=0.025, save_fn=None)[source]
plot_stats(working_dir)[source]
plot_stats2(working_dir)[source]
plot_training_history(history, par_dir)[source]
plot_wiring_weights(working_dir, with_input_blocks, with_skip_connection)[source]
rand_jitter(arr, scale=0.01)[source]
reset_plot(width_in_inches=4.5, height_in_inches=4.5)[source]
reset_style()[source]
sma(data, window=10)[source]
violin_sns(data, x, y, hue, save_fn=None, split=True, **kwargs)[source]