KaneAI - World’s First E2E Software Testing Agent.Book a Demo Join TestMu Conference

How to use parallelized method in Lemoncheesecake

Best Python code snippet using lemoncheesecake

test_dygraph_models_class_method_parallel.py

Source:test_dygraph_models_class_method_parallel.py

...71            logging.info("example #{}:".format(index))72            logging.info("label: {}".format(label_name))73            logging.info("text: {}".format(text.encode("utf-8")))74            logging.info("token_ids: {}".format(token_ids))75    def test_textcnn_parallelized(self):76        # ¶à¿¨ÔËÐÐÊ± embedding²ãµÄis_sparse²ÎÊýÐèÒªÎªFalse ÒâÎªÌÝ¶È¸üÐÂÊ±²»Ê¹ÓÃÏ¡Êè¸üÐÂ77        # ÒòÎª¶à¿¨ÑµÁ·ÔÚÊÕ¼¯ÌÝ¶ÈÊ±²»ÄÜ´¦ÀíÏ¡ÊèÌÝ¶È78        textcnn_config = {79                "num_class": TestDygraphModelsParallelized.label_encoder.size(),80                "vocab_size": TestDygraphModelsParallelized.tokenizer.size(),81                "emb_dim" : 512,82                "num_filters": 256,83                "fc_hid_dim": 512,84                "num_channels":1,85                "win_size_list": [3],86                "is_sparse": False,87                "use_cudnn": True,88                }89        run_config = {90                "model_save_path": os.path.join(TestDygraphModelsParallelized.test_output_dir, "textcnn"),91                "best_model_save_path": os.path.join(TestDygraphModelsParallelized.test_output_dir, "textcnn_best"),92                "epochs": 2,93                "batch_size": 32,94                "learning_rate": 5e-4,95                "max_seq_len": 300,96                "print_step": 200,97                "load_best_model": True,98                }99        start_time = time.time()100        class TextCNNModel(ClassificationModel):101            @model_parallelized(TestDygraphModelsParallelized.strategy)102            def build(self, **model_config):103                self.model = TextCNNClassifier(**model_config)104                self.built = True105        place = F.CUDAPlace(D.ParallelEnv().dev_id)106        with D.guard(place):107            textcnn_model = TextCNNModel()108            textcnn_model.build(**textcnn_config)109            best_acc = textcnn_model.train(110                    TestDygraphModelsParallelized.train_data, TestDygraphModelsParallelized.eval_data,111                    label_encoder=TestDygraphModelsParallelized.label_encoder,112                    **run_config)113        logging.warning("textcnn parallelized best train score: {}, cost time: {}s".format(best_acc, time.time()- start_time))114    def test_gru_parallelized(self):115        gru_config = {116                "num_class": TestDygraphModelsParallelized.label_encoder.size(),117                "vocab_size": TestDygraphModelsParallelized.tokenizer.size(),118                "emb_dim" : 512,119                "gru_dim" : 256,120                "fc_hid_dim": 512,121                "is_sparse": False,122                "bi_direction": True,123                }124        run_config = {125                "model_save_path": os.path.join(TestDygraphModelsParallelized.test_output_dir, "gru"),126                "best_model_save_path": os.path.join(TestDygraphModelsParallelized.test_output_dir, "gru_best"),127                "epochs": 2,128                "batch_size": 32,129                "max_seq_len": 300,130                "print_step": 200,131                "learning_rate": 5e-4,132                }133        start_time = time.time()134        class GRUModel(ClassificationModel):135            @model_parallelized(TestDygraphModelsParallelized.strategy)136            def build(self, **model_config):137                self.model = GRUClassifier(**model_config)138                self.built = True139        place = F.CUDAPlace(D.ParallelEnv().dev_id)140        with D.guard(place):141            gru_model = GRUModel()142            gru_model.build(**gru_config)143            best_acc = gru_model.train(144                    TestDygraphModelsParallelized.train_data, TestDygraphModelsParallelized.eval_data,145                    label_encoder=TestDygraphModelsParallelized.label_encoder,146                    **run_config)147        logging.warning("gru parallelized best train score: {}, cost time: {}s".format(best_acc, time.time()- start_time))148    def test_lstm_parallelized(self):149        lstm_config = {150                "num_class": TestDygraphModelsParallelized.label_encoder.size(),151                "vocab_size": TestDygraphModelsParallelized.tokenizer.size(),152                "emb_dim" : 512,153                "lstm_dim" : 256,154                "fc_hid_dim": 512,155                "is_sparse": False,156                "bi_direction": True,157                }158        run_config = {159                "model_save_path": os.path.join(TestDygraphModelsParallelized.test_output_dir, "lstm"),160                "best_model_save_path": os.path.join(TestDygraphModelsParallelized.test_output_dir, "lstm_best"),161                "epochs": 2,162                "batch_size": 32,163                "max_seq_len": 300,164                "print_step": 200,165                "learning_rate": 5e-4,166                }167        start_time = time.time()168        class LSTMModel(ClassificationModel):169            @model_parallelized(TestDygraphModelsParallelized.strategy)170            def build(self, **model_config):171                self.model = DynamicLSTMClassifier(**model_config)172                self.built = True173        place = F.CUDAPlace(D.ParallelEnv().dev_id)174        with D.guard(place):175            lstm_model = LSTMModel()176            lstm_model.build(**lstm_config)177            best_acc = lstm_model.train(178                    TestDygraphModelsParallelized.train_data, TestDygraphModelsParallelized.eval_data,179                    label_encoder=TestDygraphModelsParallelized.label_encoder,180                    **run_config)181        logging.warning("lstm parallelized best train score: {}, cost time: {}s".format(best_acc, time.time()- start_time))182    def test_ernie_parallelized(self):183        ernie_config = {184                "pretrain_dir_or_url": "ernie-1.0",185                "num_labels": TestDygraphModelsParallelized.label_encoder.size(),186                }187        run_config = {188                "model_save_path": os.path.join(TestDygraphModelsParallelized.test_output_dir, "ernie"),189                "best_model_save_path": os.path.join(TestDygraphModelsParallelized.test_output_dir, "ernie_best"),190                "epochs": 2,191                "batch_size": 32,192                "max_seq_len": 300,193                "print_step": 100,194                "learning_rate": 5e-5,195                "load_best_model": False,196                }197        start_time = time.time()198        class ErnieClassificationModel(ClassificationModel):199            @model_parallelized(TestDygraphModelsParallelized.strategy)200            def build(self, **model_config):201                self.model = ErnieSequenceClassificationCustomized.from_pretrained(**model_config)202                self.built = True203        place = F.CUDAPlace(D.ParallelEnv().dev_id)204        with D.guard(place):205            ernie_classification_model = ErnieClassificationModel()206            ernie_classification_model.build(**ernie_config)207            best_acc = ernie_classification_model.train(208                    TestDygraphModelsParallelized.train_data, TestDygraphModelsParallelized.eval_data,209                    label_encoder=TestDygraphModelsParallelized.label_encoder,210                    **run_config)211        logging.warning("ernie parallelized best train score: {}, cost time: {}s".format(best_acc, time.time()- start_time))212if __name__ == "__main__":213    # ÔËÐÐËùÓÐ²âÊÔÓÃÀý...

display_chart.py

Source:display_chart.py

1import matplotlib.pyplot as plt2def extract_file_column(file_name, column_index):3	column_values = [] 4	# Abrindo um arquivo em modo de leitura:5	with open("{0}.txt".format(file_name),'r') as file:6		all_file_lines = file.read().splitlines()7		for each_line in all_file_lines:8			# Separa cada linha em uma lista com9			# cada um dos trÃªs dados que cada uma contÃ©m10			data = each_line.split(' ')11			column_values.append(float(data[column_index]))12	return column_values13def extract_file_line(file_name, line_index):14	column_values = [] 15	# Abrindo um arquivo em modo de leitura:16	with open("{0}.txt".format(file_name),'r') as file:17		all_file_lines = file.read().splitlines()18		line_values = all_file_lines[line_index].split(' ')19	return line_values20def comparison_technics(number_of_threads):21	data_sizes = extract_file_column("recorded durations/{}threads_times_original".format(number_of_threads), 0)22	original_durations = extract_file_column("recorded durations/{}threads_times_original".format(number_of_threads), 1)23	optimized_durations = extract_file_column("recorded durations/{}threads_times_optimized".format(number_of_threads), 1)24	parallelized_durations = extract_file_column("recorded durations/{}threads_times_parallelized".format(number_of_threads), 1)25	optimized_and_parallelized_durations = extract_file_column("recorded durations/{}threads_times_optimized_and_parallelized".format(number_of_threads), 1)26	# Monta o grÃ¡fico com as listas obtidas:27	plt.plot(data_sizes, original_durations, color="m", label="Original", linewidth=2.0)28	plt.plot(data_sizes, optimized_durations, color="b",  label="Otimizado", linewidth=2.0)29	plt.plot(data_sizes, parallelized_durations, color="c", label="Paralelizado", linewidth=2.0)30	plt.plot(data_sizes, optimized_and_parallelized_durations, color="g",  label="Paralelizado e Otimizado", linewidth=2.0)31	plt.legend(loc='upper left', frameon=False)32	plt.xlabel('Tamanho dos Dados')33	plt.ylabel('Tempo (s)')34	plt.xlim(100, 1400)35	plt.ylim(0, 0.2)36	plt.title("GrÃ¡fico de Tempo de ExecuÃ§Ã£o por TÃ©cnica Aplicada no cÃ³digo ({} Threads)".format(number_of_threads))37	38	# Exibe o grÃ¡fico na tela:39	plt.show()40def speedup(line_index):41	data_size = extract_file_line("recorded durations/4threads_times_original", line_index)[0]42	t2_original_duration = extract_file_line("recorded durations/2threads_times_original", line_index)[1]43	t2_optimized_duration = extract_file_line("recorded durations/2threads_times_optimized", line_index)[1]44	t2_parallelized_duration = extract_file_line("recorded durations/2threads_times_parallelized", line_index)[1]45	t2_optimized_and_parallelized_duration = extract_file_line("recorded durations/2threads_times_optimized_and_parallelized", line_index)[1]46	t4_original_duration = extract_file_line("recorded durations/4threads_times_original", line_index)[1]47	t4_optimized_duration = extract_file_line("recorded durations/4threads_times_optimized", line_index)[1]48	t4_parallelized_duration = extract_file_line("recorded durations/4threads_times_parallelized", line_index)[1]49	t4_optimized_and_parallelized_duration = extract_file_line("recorded durations/4threads_times_optimized_and_parallelized", line_index)[1]50	t8_original_duration = extract_file_line("recorded durations/8threads_times_original", line_index)[1]51	t8_optimized_duration = extract_file_line("recorded durations/8threads_times_optimized", line_index)[1]52	t8_parallelized_duration = extract_file_line("recorded durations/8threads_times_parallelized", line_index)[1]53	t8_optimized_and_parallelized_duration = extract_file_line("recorded durations/8threads_times_optimized_and_parallelized", line_index)[1]54	t16_original_duration = extract_file_line("recorded durations/16threads_times_original", line_index)[1]55	t16_optimized_duration = extract_file_line("recorded durations/16threads_times_optimized", line_index)[1]56	t16_parallelized_duration = extract_file_line("recorded durations/16threads_times_parallelized", line_index)[1]57	t16_optimized_and_parallelized_duration = extract_file_line("recorded durations/16threads_times_optimized_and_parallelized", line_index)[1]58	speedup_between_original_and_parallel = [float(t2_original_duration) /float(t2_parallelized_duration) ,59                                             float(t4_original_duration) /float(t4_parallelized_duration) ,60                                             float(t8_original_duration) /float(t8_parallelized_duration) ,61                                             float(t16_original_duration)/float(t16_parallelized_duration)]62	63	speedup_between_optimizeds = [float(t2_optimized_duration) /float(t2_optimized_and_parallelized_duration) ,64                                  float(t4_optimized_duration) /float(t4_optimized_and_parallelized_duration) ,65                                  float(t8_optimized_duration) /float(t8_optimized_and_parallelized_duration) ,66                                  float(t16_optimized_duration)/float(t16_optimized_and_parallelized_duration)]67	# Monta o grÃ¡fico com as listas obtidas:68	plt.plot([2, 4, 8, 16], speedup_between_original_and_parallel, color="b",  label="Original e Paralelizado", linewidth=2.0)69	plt.plot([2, 4, 8, 16], speedup_between_optimizeds, color="c", label="Otimizado Original e Paralelizado", linewidth=2.0)70	plt.legend(loc='upper left', frameon=False)71	plt.xlabel('NÃºmero de Threads')72	plt.ylabel('Speedup')73	plt.xlim(2, 16)74	plt.ylim(1, 3)75	plt.title("GrÃ¡fico do ganho de Speedup (Tamanho do Array igual a {0}x{0})".format(str(data_size)))76	77	# Exibe o grÃ¡fico na tela:...

ray_test.py

Source:ray_test.py

1#!/usr/bin/env python2# -*- coding: UTF-8 -*-3# REF [site] >> https://github.com/ray-project/ray4import time5import ray6import tensorflow as tf7ray.init()8@ray.remote9def func():10	time.sleep(1)11	return 112def simple_example():13	results = ray.get([func.remote() for i in range(4)])14	print('Result =', results)15@ray.remote16def sqr(x):17	time.sleep(1)18	return x * x19def parallelized_map(func, args):20    return list(func.remote(arg) for arg in args)21def parallelized_map_example():22	# Call parallelized_map() on a list.23	result_ids = parallelized_map(sqr, range(1, 6))24	# Get the results.25	results = ray.get(result_ids)26	print('Result =', results)27@ray.remote28def negative(x):29	time.sleep(1)30	return x < 031def parallelized_filter(func, args):32    return list(arg for arg in args if func.remote(arg))33def parallelized_filter_example():34	# Call parallelized_filter() on a list.35	result_ids = parallelized_filter(negative, range(-5, 5))36	# Get the results.37	results = ray.get(result_ids)38	print('Result =', results)39@ray.remote40class Simulator(object):41	def __init__(self):42		self.sess = tf.Session()43		self.simple_model = tf.constant([1.0])44	def simulate(self):45		return self.sess.run(self.simple_model)46def simple_tensorflow_example():47	# Create two actors.48	simulators = [Simulator.remote() for _ in range(2)]49	# Run two simulations in parallel.50	results = ray.get([s.simulate.remote() for s in simulators])51	print(results)52def main():53	#simple_example()54	parallelized_map_example()55	#parallelized_filter_example()56	#simple_tensorflow_example()57#--------------------------------------------------------------------58if '__main__' == __name__:...

Automation Testing Tutorials

Learn to execute automation testing from scratch with LambdaTest Learning Hub. Right from setting up the prerequisites to run your first automation test, to following best practices and diving deeper into advanced test scenarios. LambdaTest Learning Hubs compile a list of step-by-step guides to help you be proficient with different test automation frameworks i.e. Selenium, Cypress, TestNG etc.