default='cv', type=str, help='Dataset for evaluation') def main(): def eval_input_fn(): return input_fn(*data[config['data']], batch_size=config['batch_size'], shuffle=False) data = inputs.load_data(config['n_examples_for_cv']) estimator = tf.estimator.Estimator(model_fn=model_fn, params=config, model_dir=config['model_dir']) for ckpt in tf.train.get_checkpoint_state( config['model_dir']).all_model_checkpoint_paths: with mu.Timer() as timer: result = estimator.evaluate(eval_input_fn, checkpoint_path=ckpt) result['data'] = config['data'] logger.info('Done in %.fs', timer.eclipsed) logger.info('\n%s\n%s%s%s\n', data, '*' * 10, result, '*' * 10) if __name__ == '__main__': tf.logging.set_verbosity(tf.logging.INFO) FLAGS = parser.parse_args() config = mu.load_config(path=None, **FLAGS.__dict__) logger.info('\n%s\n', mu.json_out(config)) main()
def main(): logger.info('\n%s\n', mu.json_out(config.state)) experiment = Experiment(config) experiment.train()