save_every = args.save_every log_every = args.log_every test_every = args.test_every primal_lr = args.primal_lr dual_lr = args.dual_lr else: continuation = False num_epochs = 1000 save_every = 100 log_every = 100 test_every = 5 primal_lr = 1e-6 dual_lr = 1e-4 # %% GENERATING DATASET ngram = randomized_ngram(7, 2, out_dim=10) ngram.show() data_loader = train_loader_MNIST() test_loader = test_loader_MNIST() sequence_loader = sequence_loader_MNIST(ngram, num_samples=20000) # %% REGULAR TRAINING (SGD) # model = Model(ngram) # model.to(DEVICE) # model.init_weights() # optimizer = torch.optim.Adam(model.primal.parameters()) # history = SGD(model, optimizer, data_loader, test_loader, num_epochs=1, log_every=50, test_every=1) # %% DUAL TRAINING
save_every = args.save_every log_every = args.log_every test_every = args.test_every primal_lr = args.primal_lr dual_lr = args.dual_lr else: continuation = False num_epochs = 1000 save_every = 100 log_every = 100 test_every = 5 primal_lr = 1e-6 dual_lr = 1e-4 # %% GENERATING DATASET ngram = randomized_ngram(10, 2, out_dim=10, min_var=2e-2) ngram.show() # %% CREATING MODEL data_loader = train_loader_MNIST() test_loader = test_loader_MNIST() sequence_loader = sequence_loader_MNIST(ngram, num_samples=20000) # %% REGULAR TRAINING (SGD) # model = Model(ngram) # model.to(DEVICE) # model.init_weights() # optimizer = torch.optim.Adam(model.primal.parameters()) # history = SGD(model, optimizer, data_loader, test_loader, num_epochs=1, log_every=50, test_every=1)
save_every = args.save_every log_every = args.log_every test_every = args.test_every primal_lr = args.primal_lr dual_lr = args.dual_lr else: continuation = False num_epochs = 1000 save_every = 100 log_every = 100 test_every = 5 primal_lr = 1e-6 dual_lr = 1e-4 # %% GENERATING DATASET ngram = randomized_ngram(3, 20, out_dim=5) data_loader = train_loader_MNIST() test_loader = test_loader_MNIST() sequence_loader = sequence_loader_MNIST(ngram, num_samples=40000) # %% REGULAR TRAINING (SGD) # model = Model(ngram) # model.to(DEVICE) # model.init_weights() # optimizer = torch.optim.Adam(model.primal.parameters()) # history = SGD(model, optimizer, data_loader, test_loader, num_epochs=1, log_every=50, test_every=1) # %% DUAL TRAINING if continuation:
num_epochs = 500 save_every = 100 log_every = 100 test_every = 2 primal_lr = 1e-6 dual_lr = 1e-4 show_dual = False predictions_on_sequences = True predictions_on_data = False ngram_data_stats = False ngram_test_stats = True loss_on_test = False # %% CREATING NGRAM ngram = randomized_ngram(3, 10, out_dim=5, min_var=1e-2) # ngram = Ngram(3) # ngram[(0, 1, 2)] = 9. # ngram[(1, 2, 3)] = 1. # ngram.norm() ngram.show() # %% GENERATING DATASET data_loader = train_loader_MNIST() test_loader = test_loader_MNIST() sequence_loader = sequence_loader_MNIST(ngram, num_samples=50000) sequence_test_loader = sequence_loader_MNIST(ngram, num_samples=10000) # %% REGULAR TRAINING (SGD) # model = Model(ngram) # model.to(DEVICE)