ensemble.py

import argparse
import pickle

import numpy as np
from tqdm import tqdm

label = open('./data/ntu/xview/val_label.pkl', 'rb')
label = np.array(pickle.load(label))
r1 = open('./work_dir/ntu_ShiftGCN_joint_xview/eval_results/best_acc.pkl', 'rb')
r1 = list(pickle.load(r1).items())
r2 = open('./work_dir/ntu_ShiftGCN_bone_xview/eval_results/best_acc.pkl', 'rb')
r2 = list(pickle.load(r2).items())
r3 = open('./work_dir/ntu_ShiftGCN_joint_motion_xview/eval_results/best_acc.pkl', 'rb')
r3 = list(pickle.load(r3).items())
r4 = open('./work_dir/ntu_ShiftGCN_bone_motion_xview/eval_results/best_acc.pkl', 'rb')
r4 = list(pickle.load(r4).items())

alpha = [0.6,0.6,0.4,0.4] # same hyperparameter is used for NTU/NTU120/NW-UCLA

right_num = total_num = right_num_5 = 0
for i in tqdm(range(len(label[0]))):
    _, l = label[:, i]
    _, r11 = r1[i]
    _, r22 = r2[i]
    _, r33 = r3[i]
    _, r44 = r4[i]
    r = r11*alpha[0] + r22 * alpha[1] + r33*alpha[2] + r44 * alpha[3]
    rank_5 = r.argsort()[-5:]
    right_num_5 += int(int(l) in rank_5)
    r = np.argmax(r)
    right_num += int(r == int(l))
    total_num += 1
acc = right_num / total_num
acc5 = right_num_5 / total_num
print('top1: ', acc)
print('top5: ', acc5)