for idy in range(seq_len) if mask[idx][idy] != 0 ] pred_label.append(pred) return pred_label seed_num = 123 random.seed(seed_num) torch.manual_seed(seed_num) np.random.seed(seed_num) if __name__ == '__main__': data = Data() data.load('./data/PoSTagger.data') predict_config_path = './predict.config' data.readConfig(predict_config_path) printParameterSummary(data) predict_instances = getDataLoader(data.infer_path, data) device = torch.device("cuda:" + data.GPU if torch.cuda.is_available() else "cpu") model = SequenceModel(data) model = torch.load(data.model_save_path) model.eval() words = pd.read_csv(data.infer_path, header=None, low_memory=False, encoding='utf-8')
else: print("Training finished. Best accuracy is {:0.4f}.".format(best_score)) seed_num = 123 random.seed(seed_num) torch.manual_seed(seed_num) np.random.seed(seed_num) if __name__ == '__main__': data = Data() train_config_path = './train.config' data.readConfig(train_config_path) data.buildDictionary() data.getPretrainedEmbedding() # print parameter summary printParameterSummary(data) # build dataloaders training_instances = getDataLoader(data.training_path, data) validation_instances = getDataLoader(data.validation_path, data) evaluation_instances = getDataLoader(data.evaluation_path, data) data.saveData() device = torch.device("cuda:"+data.GPU if torch.cuda.is_available() else "cpu") train(training_instances, validation_instances, evaluation_instances, data)