main.py

"""Main script for ADDA."""
import sys; sys.path.append("../_EXTRAS")

import matplotlib
matplotlib.use('Agg')
import torch

import argparse
import pandas as pd
import numpy as np
import random
import misc as ms
import experiments
from addons import vis
from addons import pretty_plot
import train

if __name__ == '__main__':
    # SEE IF CUDA IS AVAILABLE
    assert torch.cuda.is_available()
    print("CUDA: %s" % torch.version.cuda)
    print("Pytroch: %s" % torch.__version__)
    parser = argparse.ArgumentParser()

    parser.add_argument('-e','--expList', nargs="+", default=None) 
    parser.add_argument('-b','--borgy', default=0, type=int)
    parser.add_argument('-br','--borgy_running', default=0, type=int)
    parser.add_argument('-m','--mode', default="train")
    parser.add_argument('-rs','--reset_src', default=0, type=int)
    parser.add_argument('-rt','--reset_tgt', default=0, type=int)
    parser.add_argument('-g','--gpu', type=int)
    parser.add_argument('-s','--summary', type=int, default=0)
    parser.add_argument('-c','--configList', nargs="+",
                      default=None)
    parser.add_argument('-l','--lossList', nargs="+",
                      default=None)
    parser.add_argument('-d','--datasetList', nargs="+",
                      default=None)
    parser.add_argument('-metric','--metricList', nargs="+",
                      default=None)
    parser.add_argument('-model','--modelList', nargs="+",
                      default=None)
    parser.add_argument('-run','--plot_run', default=None, type=int)
    
    args = parser.parse_args()
    ms.set_gpu(args.gpu) 
    plot_run = args.plot_run
    
    # init random seed
    # init_random_seed(10)
    results = {}


    pp_main = pretty_plot.PrettyPlot(ratio=0.5,
                                figsize=(5,4),
                                legend_type="line",
                                yscale="linear",
                                subplots=(1, 1),
                                shareRowLabel=True)

    
        

    #for exp_name in args.expList:
    
    #exp_name = 'indianSrc2indianTgt'
    exp_name = 'paviaSrc2paviaTgt'
    #exp_name = 'shSrc2shTgt'
    
    #exp_name = 'indianTgt'
    #exp_name = 'paviaTgt'
    #exp_name = 'shTgt'
    
    #exp_name = 'indianSrc_indianTgt'
    #exp_name = 'paviaSrc_paviaTgt'
    #exp_name = 'shSrc_shTgt'
    
    N_Runs = 10
    if plot_run is not None:
        N_Runs = 1
    
    for run in range(N_Runs):
        if plot_run is not None:
            run = plot_run
        exp_dict = experiments.get_experiment_dict(args, exp_name)
        exp_dict["reset_src"] = args.reset_src
        exp_dict["reset_tgt"] = args.reset_tgt
        exp_dict['run'] = run
        
        ##SET SEED
        # np.random.seed(100)
        # random.seed(100)
        # torch.manual_seed(100) 
        # torch.cuda.manual_seed_all(100)

        history = ms.load_history(exp_dict)

        # Main options
        if args.mode == "test_model":
            exp_dict['sample_already'] = True
            results[run] = ms.test_latest_model_bing(exp_dict, verbose=1)


        elif args.mode == "train":
            train.train(exp_dict)


        if args.mode == "copy_models":
            results[exp_name] = ms.copy_models(exp_dict, path_dst="{}/".format(exp_name))


        # MISC
        if args.mode == "plot_src":

            src_losses = np.array(pd.DataFrame(history["src_train"])["loss"])
            src_epochs = np.array(pd.DataFrame(history["src_train"])["epoch"])

            pp_main.add_yxList(y_vals=src_losses[1:101], 
                               x_vals=src_epochs[1:101], 
                               label=exp_name.split("2")[0].upper().replace("BIG",""),
                               converged=None)

        if args.mode == "plot_tgt":

            tgt_acc = np.array(pd.DataFrame(history["tgt_train"])["acc_tgt"])
            src_epochs = np.array(pd.DataFrame(history["tgt_train"])["epoch"])

            pp_main.add_yxList(y_vals=tgt_acc[1:101], 
                               x_vals=src_epochs[1:101], 
                               label=exp_name.split("2")[1].upper().replace("BIG",""),
                               converged=None)
            # vis.visEmbed(exp_dict)



        if args.mode == "vis":
            vis.visEmbed(exp_dict)


        elif args.mode == "summary":

            summary = pd.DataFrame(history["tgt_train"][1:])["acc_tgt"]
            print(summary.describe())

        elif args.mode == "acc_tgt":

            summary = pd.DataFrame(history["tgt_train"][1:200])["acc_tgt"]
            print(summary)

        elif args.mode == "max":
            try:
                summary = pd.DataFrame(history["tgt_train"][1:])["acc_tgt"]
                results[exp_name] = summary.max()
            except:
                print("{} skipped...".format(exp_name))
        
    print(pd.Series(results))
    # Train Source
    if args.mode == "plot_src":
        pp_main.plot(ylabel="Triplet Loss", 
                     xlabel="Epochs",
                     yscale="log")

        path = exp_dict["summary_path"]
        pp_main.fig.tight_layout(rect=[0, 0.03, 1, 0.95])
        figName = "%s/png_plots/SRC_%s.png" % (path, exp_name)
        ms.create_dirs(figName)
        pp_main.fig.savefig(figName)

        pp_main.fig.tight_layout()
        pp_main.fig.suptitle("")

        figName = "%s/pdf_plots/SRC_%s.pdf" % (path, exp_name)
        ms.create_dirs(figName)
        pp_main.fig.savefig(figName, dpi = 600)

        print("saved {}".format(figName))
          
    if args.mode == "plot_tgt":
        pp_main.plot(ylabel="Classifcation Accuracy", 
                     xlabel="Epochs",
                     yscale="log")

        path = exp_dict["summary_path"]
        pp_main.fig.tight_layout(rect=[0, 0.03, 1, 0.95])
        figName = "%s/png_plots/TGT_%s.png" % (path, exp_name)
        ms.create_dirs(figName)
        pp_main.fig.savefig(figName)

        pp_main.fig.tight_layout()
        pp_main.fig.suptitle("")

        figName = "%s/pdf_plots/TGT_%s.pdf" % (path, exp_name)
        ms.create_dirs(figName)
        pp_main.fig.savefig(figName, dpi = 600)

        print("saved {}".format(figName))
    ####################################################
    if args.mode == "test_model":
        OA = 0.
        AA = 0.
        Kappa = 0.
        for run in range(N_Runs):
            OA = OA + results[run]['OA']
            AA = AA + results[run]['AA']
            Kappa = Kappa + results[run]['Kappa']
        print("====================="
            "\nOvearll Accuracy {}\n"
            "=====================".format(OA/N_Runs))
        print("====================="
            "\nAverage Accuracy {}\n"
            "=====================".format(AA/N_Runs))
        print("====================="
            "\nKappa {}\n"
            "=====================".format(Kappa/N_Runs))
        result={}
        result['OA'] = [OA/N_Runs]
        result['AA'] = [AA/N_Runs]
        result['Kappa'] = [Kappa/N_Runs]
        ms.save_json('results/'+exp_name+'_center_{}_disc_{}.json'.format(exp_dict["options"]["center"],exp_dict["options"]["disc"]), result)