Unable to predict using single test image with no ground truth image

[purijs]

My export params:

parser.add_argument("--input_dir", default="./data/combined/", help="path to folder containing images")
parser.add_argument("--mode", default="export", choices=["train", "test", "export"])
parser.add_argument("--output_dir", default="./exports", help="where to put output files")
parser.add_argument("--seed", type=int, default=121)
parser.add_argument("--checkpoint", default='./output', help="directory with checkpoint to resume training from or use for testing")

This is what is exported:

Testing params:

parser = argparse.ArgumentParser()
parser.add_argument("--model_dir", default="./exports/", help="directory containing exported model")
parser.add_argument("--input_file", default="./test.png", help="input PNG image file")
parser.add_argument("--output_file", default="./output.png", help="output PNG image file")
a = parser.parse_args([])

Error: It wont work for "/export" also instead of "/export/checkpoint"