import os import sys sys.path.append(os.path.abspath(os.path.dirname(__file__) + '/' + '..')) from models import gru_simple from helpers import checkpoint # Get the review summary file review_summary_file = 'extracted_data/review_summary.csv' # Initialize Checkpointer to ensure checkpointing checkpointer = checkpoint.Checkpointer('simple', 'gru', 'Attention') checkpointer.steps_per_checkpoint(1000) checkpointer.steps_per_prediction(1000) # Do using GRU cell - with attention mechanism out_file = 'result/simple/gru/attention.csv' checkpointer.set_result_location(out_file) gru_net = gru_simple.GruSimple(review_summary_file, checkpointer, attention=True) gru_net.set_parameters(train_batch_size=128, test_batch_size=128, memory_dim=128, learning_rate=0.05) gru_net.begin_session() gru_net.form_model_graph() gru_net.fit() gru_net.predict() gru_net.store_test_predictions()
import os import sys sys.path.append(os.path.abspath(os.path.dirname(__file__) + '/' + '..')) from models import lstm_simple from helpers import checkpoint # Get the review summary file review_summary_file = 'extracted_data/review_summary.csv' # Initialize Checkpointer to ensure checkpointing checkpointer = checkpoint.Checkpointer('simple', 'lstm', 'Attention') checkpointer.steps_per_checkpoint(1000) checkpointer.steps_per_prediction(1000) # Do using LSTM cell - with attention mechanism out_file = 'result/simple/lstm/attention.csv' checkpointer.set_result_location(out_file) lstm_net = lstm_simple.LstmSimple(review_summary_file, checkpointer, attention=True) lstm_net.set_parameters(train_batch_size=128, test_batch_size=128, memory_dim=128, learning_rate=0.05) lstm_net.begin_session() lstm_net.form_model_graph() lstm_net.fit() lstm_net.predict() lstm_net.store_test_predictions()
import os import sys sys.path.append(os.path.abspath(os.path.dirname(__file__) + '/' + '..')) from models import gru_stacked_bidirectional from helpers import checkpoint # Get the review summary file review_summary_file = 'extracted_data/review_summary.csv' # Initialize Checkpointer to ensure checkpointing checkpointer = checkpoint.Checkpointer('stackedBidirectional', 'gru', 'Attention') checkpointer.steps_per_checkpoint(1000) checkpointer.steps_per_prediction(1000) # Do using GRU cell - without attention mechanism out_file = 'result/stacked_bidirectional/gru/attention.csv' checkpointer.set_result_location(out_file) gru_net = gru_stacked_bidirectional.GruStackedBidirectional( review_summary_file, checkpointer, attention=True, num_layers=2) gru_net.set_parameters(train_batch_size=128, test_batch_size=128, memory_dim=128, learning_rate=0.05) gru_net.begin_session() gru_net.form_model_graph() gru_net.fit() gru_net.predict() gru_net.store_test_predictions()
import os import sys sys.path.append(os.path.abspath(os.path.dirname(__file__) + '/' + '..')) from models import lstm_bidirectional from helpers import checkpoint # Get the review summary file review_summary_file = 'extracted_data/review_summary.csv' # Initialize Checkpointer to ensure checkpointing checkpointer = checkpoint.Checkpointer('bidirectional', 'lstm', 'Attention') checkpointer.steps_per_checkpoint(1000) checkpointer.steps_per_prediction(1000) # Do using GRU cell - without attention mechanism out_file = 'result/bidirectional/lstm/attention.csv' checkpointer.set_result_location(out_file) lstm_net = lstm_bidirectional.LstmBidirectional(review_summary_file, checkpointer, attention=True) lstm_net.set_parameters(train_batch_size=128, test_batch_size=128, memory_dim=128, learning_rate=0.05) lstm_net.begin_session() lstm_net.form_model_graph() lstm_net.fit() lstm_net.predict() lstm_net.store_test_predictions()
from models import lstm_stacked_simple from helpers import checkpoint # Get the review summary file review_summary_file = 'extracted_data/review_summary.csv' # Initialize Checkpointer to ensure checkpointing checkpointer = checkpoint.Checkpointer('stackedSimple', 'lstm', 'Attention') checkpointer.steps_per_checkpoint(500) checkpointer.steps_per_prediction(2000) # Do using GRU cell - without attention mechanism out_file = 'result/stacked_simple/lstm/attention.csv' checkpointer.set_result_location(out_file) lstm_net = lstm_stacked_simple.NeuralNet(review_summary_file, checkpointer, attention=True) lstm_net.set_parameters(train_batch_size=5, test_batch_size=25, memory_dim=50, learning_rate=0.05) lstm_net.form_model_graph(num_layers=2) lstm_net.fit() lstm_net.predict() lstm_net.store_test_predictions() lstm_net.close_session()
import os import sys sys.path.append(os.path.abspath(os.path.dirname(__file__) + '/' + '..')) from models import gru_bidirectional from helpers import checkpoint # Get the review summary file review_summary_file = 'extracted_data/review_summary.csv' # Initialize Checkpointer to ensure checkpointing checkpointer = checkpoint.Checkpointer('bidirectional', 'gru', 'noAttention') checkpointer.steps_per_checkpoint(1000) checkpointer.steps_per_prediction(1000) # Do using GRU cell - without attention mechanism out_file = 'result/bidirectional/gru/no_attention.csv' checkpointer.set_result_location(out_file) gru_net = gru_bidirectional.GruBidirectional(review_summary_file, checkpointer) gru_net.set_parameters(train_batch_size=128, test_batch_size=128, memory_dim=128, learning_rate=0.05) gru_net.begin_session() gru_net.form_model_graph() gru_net.fit() gru_net.predict() gru_net.store_test_predictions()
import os import sys sys.path.append(os.path.abspath(os.path.dirname(__file__) + '/' + '..')) from models import gru_stacked_simple from helpers import checkpoint # Get the review summary file review_summary_file = 'extracted_data/review_summary.csv' # Initialize Checkpointer to ensure checkpointing checkpointer = checkpoint.Checkpointer('stackedSimple', 'gru', 'noAttention') checkpointer.steps_per_checkpoint(1000) checkpointer.steps_per_prediction(1000) # Do using GRU cell - without attention mechanism out_file = 'result/stacked_simple/gru/no_attention.csv' checkpointer.set_result_location(out_file) gru_net = gru_stacked_simple.GruStackedSimple(review_summary_file, checkpointer, num_layers=2) gru_net.set_parameters(train_batch_size=128, test_batch_size=128, memory_dim=128, learning_rate=0.05) gru_net.begin_session() gru_net.form_model_graph() gru_net.fit() gru_net.predict() gru_net.store_test_predictions()
import os import sys sys.path.append(os.path.abspath(os.path.dirname(__file__) + '/' + '..')) from models import lstm_stacked_bidirectional from helpers import checkpoint # Get the review summary file review_summary_file = 'extracted_data/review_summary.csv' # Initialize Checkpointer to ensure checkpointing checkpointer = checkpoint.Checkpointer('stackedBidirectional', 'lstm', 'noAttention') checkpointer.steps_per_checkpoint(1000) checkpointer.steps_per_prediction(1000) # Do using GRU cell - without attention mechanism out_file = 'result/stacked_bidirectional/lstm/no_attention.csv' checkpointer.set_result_location(out_file) lstm_net = lstm_stacked_bidirectional.LstmStackedBidirectional( review_summary_file, checkpointer, num_layers=2) lstm_net.set_parameters(train_batch_size=128, test_batch_size=128, memory_dim=128, learning_rate=0.05) lstm_net.begin_session() lstm_net.form_model_graph() lstm_net.fit() lstm_net.predict() lstm_net.store_test_predictions()