output_file=folder_images +
"/training_weights.avi",
fps=2)
# pf.create_video_from_images(video_fotograms_folder3,output_file = "./training_Variational_weights.avi", fps = 2)
pf.create_video_from_images(video_fotograms_folder4,
output_file=folder_images +
"/training_Variational_weights.avi",
fps=2)
"""
###################### SAVE MODEL ####################
"""
myGeneralVBModel.save(folder_model + "model_parameters_epoch:%i.pk" % i)
"""
##################### STATIC PLOTS ######################
"""
# Set it in no training
myGeneralVBModel.eval()
if (plot_predictions):
####### PLOT THE LEARNT FUNCTION ############
pf.create_Bayesian_analysis_charts(myGeneralVBModel, X_data_tr, Y_data_tr,
X_data_val, Y_data_val, tr_data_loss,
val_data_loss, KL_loss, xgrid_real_func,
ygrid_real_func, folder_images)
if (plot_weights):
####### PLOT ANALYSIS OF THE WEIGHTS ###########
pf.plot_weights_network(myGeneralVBModel, folder_images)
pf.plot_Variational_weights_network(myGeneralVBModel, folder_images)
# pf.create_video_from_images(video_fotograms_folder3,output_file = "./training_Variational_weights.avi", fps = 2)
pf.create_video_from_images(video_fotograms_folder4,output_file = folder_images + "/training_Variational_weights.avi", fps = 2)
"""
###################### SAVE MODEL ####################
"""
myHalfBayesianMLP.save(folder_model + "model_parameters_epoch:%i.pk"%i)
"""
##################### STATIC PLOTS ######################
"""
# Set it in no training
myHalfBayesianMLP.eval()
if (plot_predictions):
####### PLOT THE LEARNT FUNCTION ############
pf.create_Bayesian_analysis_charts(myHalfBayesianMLP,
X_data_tr, Y_data_tr, X_data_val, Y_data_val,
tr_loss, val_loss, KL_loss,
xgrid_real_func, ygrid_real_func,
folder_images)
if (plot_weights):
####### PLOT ANALYSIS OF THE WEIGHTS ###########
pf.plot_weights_network(myHalfBayesianMLP, folder_images)
pf.plot_Variational_weights_network(myHalfBayesianMLP, folder_images)
fps=2)
"""
###################### SAVE MODEL ####################
"""
myBasicMLP.save(folder_model + "model_parameters_epoch:%i.pk" % i)
"""
##################### STATIC PLOTS ######################
"""
# Set it in no training
myBasicMLP.eval()
if (plot_predictions):
####### PLOT THE LEARNT FUNCTION ############
x_grid = np.linspace(
np.min([X_data_tr]) - 1,
np.max([X_data_val]) + 10, 10000)
y_grid = myBasicMLP.predict(
torch.tensor(x_grid.reshape(-1, 1), device=device,
dtype=dtype)).detach().numpy()
pf.plot_learnt_function(X_data_tr, Y_data_tr, X_data_val, Y_data_val,
x_grid, y_grid, cf_a, folder_images)
if (plot_evolution_loss):
####### PLOT THE EVOLUTION OF RMSE ############
pf.plot_evolution_RMSE(tr_loss, val_loss, cf_a, folder_images)
if (plot_weights):
####### PLOT ANALYSIS OF THE WEIGHTS ###########
pf.plot_weights_network(myBasicMLP, folder_images)
if(create_video_training): #
pf.create_video_from_images(video_fotograms_folder_weights,output_file =folder_images + "/training_weights.avi", fps = 2)
pf.create_video_from_images(video_fotograms_folder_training,output_file = folder_images + "/training_Variational_weights.avi", fps = 2)
"""
###################### SAVE MODEL ####################
"""
myGeneralVBModel.save(folder_model + "model_parameters_epoch:%i.pk"%i)
"""
##################### STATIC PLOTS ######################
"""
# Set it in no training
myGeneralVBModel.eval()
if (plot_predictions):
####### PLOT THE LEARNT FUNCTION ############
pf.create_Bayesian_analysis_charts(myGeneralVBModel,
X_data_tr, Y_data_tr, X_data_val, Y_data_val,
tr_data_loss, val_data_loss, KL_loss,final_loss_tr,final_loss_val,
xgrid_real_func, ygrid_real_func,
folder_images)
if (plot_weights):
####### PLOT ANALYSIS OF THE WEIGHTS ###########
pf.plot_weights_network(myGeneralVBModel, folder_images)
pf.plot_Variational_weights_network(myGeneralVBModel, folder_images)
pf.create_video_from_images(video_fotograms_folder,output_file = "./training_loss.avi", fps = 2)
pf.create_video_from_images(video_fotograms_folder2,output_file = "./training_weights.avi", fps = 2)
"""
###################### SAVE MODEL ####################
"""
myBasicMLP.save(folder_model + "model_parameters_epoch:%i.pk"%i)
"""
##################### STATIC PLOTS ######################
"""
# Set it in no training
myBasicMLP.eval()
if (plot_predictions):
####### PLOT THE LEARNT FUNCTION ############
x_grid = np.linspace(np.min([X_data_tr]) -1, np.max([X_data_val]) +10, 10000)
y_grid = myBasicMLP.predict(torch.tensor(x_grid.reshape(-1,1),device=device, dtype=dtype)).detach().numpy()
pf.plot_learnt_function(X_data_tr, Y_data_tr, X_data_val, Y_data_val,
x_grid, y_grid, cf_a,
folder_images)
if(plot_evolution_loss):
####### PLOT THE EVOLUTION OF RMSE ############
pf.plot_evolution_RMSE(tr_loss, val_loss, cf_a,folder_images)
if (plot_weights):
####### PLOT ANALYSIS OF THE WEIGHTS ###########
pf.plot_weights_network(myBasicMLP, folder_images)