pickle.dump(colors, fp) names = [] img = [] for image in listing: x = [] X = LoadBatches.getImageArr(path1 + image, img_rows, img_cols, dp) # print(X.shape) X.reshape(img_rows, img_cols, img_channels) pr = model.predict(np.array([X]))[0] pr = np.array(pr) pr = pr.reshape((img_rows, img_cols)) # .argmax(axis=2) print(pr) # break seg_img = np.zeros((img_rows, img_cols, 1)) for i in range(img_rows): for j in range(img_cols): if pr[i][j] < 0.5: seg_img[i][j] = 0 print('absent') else: seg_img[i][j] = 255 print('present') seg_img = cv2.resize(seg_img, (101, 101)) cv2.imwrite(path2 + image, seg_img) name = ''.join(list(image)[:-4]) img.append(seg_img) names.append(name) #break make_submission(img, names, fast=False, path='result1.csv')
with open('/home/titanx/Desktop/Mainak/TGS SALT/dp', 'rb') as f: dp = pickle.load(f) names = [] img = [] for image in listing: x = [] X = LoadBatches.getImageArr(path1 + image, img_rows, img_cols, dp) # print(X.shape) X.reshape(img_rows, img_cols, img_channels) pr = model.predict(np.array([X]))[0] pr = np.array(pr) pr = pr.reshape((img_rows, img_cols)) # .argmax(axis=2) print(pr) seg_img = np.zeros((img_rows, img_cols, 1)) for i in range(img_rows): for j in range(img_cols): if pr[i][j] < 0.5: seg_img[i][j] = 0 print('absent') else: seg_img[i][j] = 255 print('present') seg_img = cv2.resize(seg_img, (101, 101)) cv2.imwrite(path2 + image, seg_img) #break name = ''.join(list(image)[:-4]) img.append(seg_img) names.append(name) #break make_submission(img, names, fast=False, path='result4_selufull.csv')
collate_fn=tta_collate if args.use_tta else default_collate, shuffle=False) model = BaseModel(args) model.init_model() model.load_trained_model() pred = model.eval(test_dataloader) # after softmax array pickle.dump({'pred': pred}, open(pkl_name, 'wb')) preds_test += pred preds_test /= len(args.ensemble_exp) # generate csv print('generate csv ...') for t in [0.29, 0.30, 0.31, 0.32, 0.33, 0.4, 0.45]: make_submission((preds_test > t).astype(np.uint8), test['names'], path='{}_{:.2f}_submission.csv'.format(csv_name, t)) print('generate {}_submission.csv'.format(args.exp_name)) if args.vis: print('vis mask ...') OUT_1 = os.path.join('Visualize', args.exp_name + '_eval') if not os.path.exists(OUT_1): os.makedirs(OUT_1) for i, d in enumerate(test['names']): name = d tmp = preds_test[i] save_img(tmp > 0.5, os.path.join(OUT_1, name + '_pred.jpg'))
from rlen import make_submission if __name__ == '__main__': import pickle pkl_path = 'train.pkl' data = pickle.load(open(pkl_path, 'rb')) masks, name = data['masks'], data['names'] make_submission(masks, name, fast=True, path='fast_submission.csv') make_submission(masks, name, fast=False, path='slow_submission.csv')