-
Notifications
You must be signed in to change notification settings - Fork 0
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
Showing
53 changed files
with
3,493 additions
and
5 deletions.
There are no files selected for viewing
Binary file added
BIN
+36.4 KB
benchmark/Glove-Twitter/Accuracy versus Epoches (Execution Mode - Original).png
.png){kind=link}
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,158 @@ | ||
| import numpy as np | ||
| from classification_func import classification, classification_lr | ||
| import update_embed_mt2st_switch | ||
| import update_embed_mt2st_dimish | ||
| import update_embed_mt | ||
| from data_preprocessing import preprocess_dataset | ||
| from eval_func import sim_loss_cal, word_sim_dir,word_vec_file, read_word_vectors | ||
| from func_plot import plot2D_accuracy, plot2D_loss, plot2D2 | ||
|
|
||
| def original_embedding(word_model): | ||
| emb_table = [] | ||
| for i, word in enumerate(word_list): | ||
| if word in word_model: | ||
| emb_table.append(word_model[word]) | ||
| else: | ||
| emb_table.append([0]*emb_dim) | ||
| emb_table = np.array(emb_table) | ||
| return emb_table | ||
|
|
||
| n_class, text_final, word_index, label, emb_dim, word_list, label_encoded, train_pad_encoded,test_pad_encoded,label_train_encoded,label_test_encoded,text_pad_encoded=preprocess_dataset() | ||
|
|
||
| mode = "mt2st_dimish" | ||
| update_modes = "null" | ||
|
|
||
| if(mode=="mt"): | ||
| update_modes = update_embed_mt | ||
| elif(mode=="mt2st_dimish"): | ||
| update_modes = update_embed_mt2st_dimish | ||
| elif(mode=="mt2st_switch"): | ||
| update_modes = update_embed_mt2st_switch | ||
|
|
||
| # Pretrained embedding Glove-Twitter | ||
| import gensim.downloader as api | ||
| wv_emb_twitter_100 = api.load("glove-twitter-100") | ||
|
|
||
| label = "Original Methods" | ||
| print(label+"\n") | ||
| wv_twitter_100_emb = original_embedding(wv_emb_twitter_100) | ||
| losses,train_accs = classification_lr(wv_twitter_100_emb,0.05, n_class, emb_dim, train_pad_encoded, label_train_encoded, test_pad_encoded, label_test_encoded) | ||
|
|
||
| epoches = [i for i in range(1, 26)] | ||
| plot2D_loss(epoches, losses, "Epoches", "Loss", | ||
| f"Loss versus Epoches (Execution Mode - Original)", | ||
| f"Execution Mode - Original") | ||
|
|
||
| plot2D_accuracy(epoches, train_accs, "Epoches", "Accuracy", | ||
| f"Accuracy versus Epoches (Execution Mode - Original)", | ||
| f"Execution Mode - Original") | ||
|
|
||
| print("losses:\n",losses) | ||
| print("train_accs:\n",train_accs) | ||
|
|
||
| print("UPDATE0 Methods\n") | ||
| original_wv_twitter_100_prediction = classification(wv_twitter_100_emb, n_class, 256) | ||
| if(update_modes!="mt2st_switch"): | ||
| losses, train_accs = update_modes.update_embedding0(wv_twitter_100_emb, n_class, emb_dim, word_list, train_pad_encoded, label_train_encoded, test_pad_encoded, | ||
| label_test_encoded, 0.95) | ||
| epoches = [i for i in range(1, 26)] | ||
| plot2D_loss(epoches, losses, "Epoches", "Loss", | ||
| f"Loss versus Epoches (Execution Mode - {mode} Update Mode - Setting0)", | ||
| f"Execution Mode - {mode} Update Mode - Setting0") | ||
|
|
||
| plot2D_accuracy(epoches, train_accs, "Epoches", "Accuracy", | ||
| f"Accuracy versus Epoches (Execution Mode - {mode} Update Mode - Setting0)", | ||
| f"Execution Mode - {mode} Update Mode - Setting0") | ||
| else: | ||
| losses_previous, losses_after, train_accs = update_modes.update_embedding0(wv_twitter_100_emb, n_class, emb_dim, word_list, | ||
| train_pad_encoded, label_train_encoded, test_pad_encoded, | ||
| label_test_encoded) | ||
| epoches1 = [i for i in range(1,16)] | ||
| epoches2 = [i for i in range(16,26)] | ||
|
|
||
| plot2D2(epoches1, losses_previous, epoches2, losses_after, "Epoches", "Accuracy", | ||
| f"Accuracy versus Epoches (Mode - {mode} Update Mode - Setting0)", f"{mode} (before switch) - Setting0", | ||
| f"{mode} (after switch) - Setting1") | ||
| plot2D2(epoches1, losses_previous, epoches2, losses_after, "Epoches", "Accuracy", | ||
| f"Accuracy versus Epoches (Mode - {mode} Update Mode - Setting0)", f"{mode} (before switch) - Setting0", | ||
| f"{mode} (after switch) - Setting0") | ||
| print("UPDATE1 Methods\n") | ||
| if(update_modes!="mt2st_switch"): | ||
| losses, train_accs = update_modes.update_embedding1(wv_twitter_100_emb, n_class, emb_dim, word_list, train_pad_encoded, label_train_encoded, test_pad_encoded, | ||
| label_test_encoded, 0.95) | ||
| epoches = [i for i in range(1,26)] | ||
| plot2D_loss(epoches, losses, "Epoches", "Loss", | ||
| f"Loss versus Epoches (Execution Mode - {mode} Update Mode - Setting1)", | ||
| f"Execution Mode - {mode} Update Mode - Setting1") | ||
|
|
||
| plot2D_accuracy(epoches, train_accs, "Epoches", "Accuracy", | ||
| f"Accuracy versus Epoches (Execution Mode - {mode} Update Mode - Setting1)", | ||
| f"Execution Mode - {mode} Update Mode - Setting1") | ||
| else: | ||
| losses_previous, losses_after, train_accs = update_modes.update_embedding1(wv_twitter_100_emb, n_class, emb_dim, word_list, | ||
| train_pad_encoded, label_train_encoded, test_pad_encoded, | ||
| label_test_encoded) | ||
|
|
||
| epoches1 = [i for i in range(1,16)] | ||
| epoches2 = [i for i in range(16,26)] | ||
|
|
||
| plot2D2(epoches1, losses_previous, epoches2, losses_after, "Epoches", "Accuracy", | ||
| f"Accuracy versus Epoches (Mode - {mode} Update Mode - Setting1)", f"{mode} (before switch) - Setting1", | ||
| f"{mode} (after switch) - Setting1") | ||
| plot2D2(epoches1, losses_previous, epoches2, losses_after, "Epoches", "Accuracy", | ||
| f"Accuracy versus Epoches (Mode - {mode} Update Mode - Setting1)", f"{mode} (before switch) - Setting1", | ||
| f"{mode} (after switch) - Setting1") | ||
|
|
||
| print("UPDATE2 Methods\n") | ||
| if(update_modes!="mt2st_switch"): | ||
| losses, train_accs = update_modes.update_embedding2(wv_twitter_100_emb, n_class, emb_dim, word_list, train_pad_encoded, label_train_encoded, test_pad_encoded, | ||
| label_test_encoded, 0.95) | ||
| epoches = [i for i in range(1,26)] | ||
| plot2D_loss(epoches, losses, "Epoches", "Loss", | ||
| f"Loss versus Epoches (Execution Mode - {mode} Update Mode - Setting2)", | ||
| f"Execution Mode - {mode} Update Mode - Setting2") | ||
|
|
||
| plot2D_accuracy(epoches, train_accs, "Epoches", "Accuracy", | ||
| f"Accuracy versus Epoches (Execution Mode - {mode} Update Mode - Setting2)", | ||
| f"Execution Mode - {mode} Update Mode - Setting2") | ||
| else: | ||
| losses_previous, losses_after, train_accs = update_modes.update_embedding2(wv_twitter_100_emb, n_class, emb_dim, word_list, | ||
| train_pad_encoded, label_train_encoded, test_pad_encoded, | ||
| label_test_encoded) | ||
|
|
||
| epoches1 = [i for i in range(1,16)] | ||
| epoches2 = [i for i in range(16,26)] | ||
|
|
||
| plot2D2(epoches1, losses_previous, epoches2, losses_after, "Epoches", "Accuracy", | ||
| f"Accuracy versus Epoches (Mode - {mode} Update Mode - Setting2)", f"{mode} (before switch) - Setting2", | ||
| f"{mode} (after switch) - Setting2") | ||
| plot2D2(epoches1, losses_previous, epoches2, losses_after, "Epoches", "Accuracy", | ||
| f"Accuracy versus Epoches (Mode - {mode} Update Mode - Setting2)", f"{mode} (before switch) - Setting2", | ||
| f"{mode} (after switch) - Setting2") | ||
|
|
||
|
|
||
| print("UPDATE3 Methods\n") | ||
| if(update_modes!="mt2st_switch"): | ||
| losses, train_accs = update_modes.update_embedding3(wv_twitter_100_emb, n_class, emb_dim, word_list, train_pad_encoded, label_train_encoded, test_pad_encoded, | ||
| label_test_encoded, 0.95) | ||
| epoches = [i for i in range(1,26)] | ||
| plot2D_loss(epoches, losses, "Epoches", "Loss", | ||
| f"Loss versus Epoches (Execution Mode - {mode} Update Mode - Setting3)", | ||
| f"Execution Mode - {mode} Update Mode - Setting3") | ||
|
|
||
| plot2D_accuracy(epoches, train_accs, "Epoches", "Accuracy", | ||
| f"Accuracy versus Epoches (Execution Mode - {mode} Update Mode - Setting3)", | ||
| f"Execution Mode - {mode} Update Mode - Setting3") | ||
| else: | ||
| losses_previous, losses_after, train_accs = update_modes.update_embedding3(wv_twitter_100_emb, n_class, emb_dim, word_list, | ||
| train_pad_encoded, label_train_encoded, test_pad_encoded, | ||
| label_test_encoded) | ||
|
|
||
| epoches1 = [i for i in range(1,16)] | ||
| epoches2 = [i for i in range(16,26)] | ||
| plot2D2(epoches1, losses_previous, epoches2, losses_after, "Epoches", "Accuracy", | ||
| f"Accuracy versus Epoches (Mode - {mode} Update Mode - Setting3)", f"{mode} (before switch) - Setting3", | ||
| f"{mode} (after switch) - Setting3") | ||
| plot2D2(epoches1, losses_previous, epoches2, losses_after, "Epoches", "Accuracy", | ||
| f"Accuracy versus Epoches (Mode - {mode} Update Mode - Setting3)", f"{mode} (before switch) - Setting3", | ||
| f"{mode} (after switch) - Setting3") |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,162 @@ | ||
| import numpy as np | ||
| from classification_func import classification, classification_lr | ||
| import update_embed_mt2st_switch | ||
| import update_embed_mt2st_dimish | ||
| import update_embed_mt | ||
| from data_preprocessing import preprocess_dataset | ||
| from eval_func import sim_loss_cal, word_sim_dir,word_vec_file, read_word_vectors | ||
| from func_plot import plot2D_accuracy, plot2D_loss, plot2D2 | ||
|
|
||
| def original_embedding(word_model): | ||
| emb_table = [] | ||
| for i, word in enumerate(word_list): | ||
| if word in word_model: | ||
| emb_table.append(word_model[word]) | ||
| else: | ||
| emb_table.append([0]*emb_dim) | ||
| emb_table = np.array(emb_table) | ||
| return emb_table | ||
|
|
||
| n_class, text_final, word_index, label, emb_dim, word_list, label_encoded, train_pad_encoded,test_pad_encoded,label_train_encoded,label_test_encoded,text_pad_encoded=preprocess_dataset() | ||
|
|
||
| mode = "mt" | ||
| update_modes = "null" | ||
|
|
||
| if(mode=="mt"): | ||
| update_modes = update_embed_mt | ||
| elif(mode=="mt2st_dimish"): | ||
| update_modes = update_embed_mt2st_dimish | ||
| elif(mode=="mt2st_switch"): | ||
| update_modes = update_embed_mt2st_switch | ||
|
|
||
|
|
||
|
|
||
| # Pretrained embedding Glove-Twitter | ||
| import gensim.downloader as api | ||
| wv_emb_twitter_100 = api.load("glove-twitter-100") | ||
|
|
||
|
|
||
|
|
||
| label = "Original Methods" | ||
| print(label+"\n") | ||
| wv_twitter_100_emb = original_embedding(wv_emb_twitter_100) | ||
| losses,train_accs = classification_lr(wv_twitter_100_emb,0.05, n_class, emb_dim, train_pad_encoded, label_train_encoded, test_pad_encoded, label_test_encoded) | ||
|
|
||
| epoches = [i for i in range(1, 26)] | ||
| plot2D_loss(epoches, losses, "Epoches", "Loss", | ||
| f"Loss versus Epoches (Execution Mode - Original)", | ||
| f"Execution Mode - Original") | ||
|
|
||
| plot2D_accuracy(epoches, train_accs, "Epoches", "Accuracy", | ||
| f"Accuracy versus Epoches (Execution Mode - Original)", | ||
| f"Execution Mode - Original") | ||
|
|
||
| print("losses:\n",losses) | ||
| print("train_accs:\n",train_accs) | ||
|
|
||
| print("UPDATE0 Methods\n") | ||
| original_wv_twitter_100_prediction = classification(wv_twitter_100_emb, n_class, 256) | ||
| if(update_modes!="mt2st_switch"): | ||
| losses, train_accs = update_modes.update_embedding0(wv_twitter_100_emb, n_class, emb_dim, word_list, train_pad_encoded, label_train_encoded, test_pad_encoded, | ||
| label_test_encoded) | ||
| epoches = [i for i in range(1, 26)] | ||
| plot2D_loss(epoches, losses, "Epoches", "Loss", | ||
| f"Loss versus Epoches (Execution Mode - {mode} Update Mode - Setting0)", | ||
| f"Execution Mode - {mode} Update Mode - Setting0") | ||
|
|
||
| plot2D_accuracy(epoches, train_accs, "Epoches", "Accuracy", | ||
| f"Accuracy versus Epoches (Execution Mode - {mode} Update Mode - Setting0)", | ||
| f"Execution Mode - {mode} Update Mode - Setting0") | ||
| else: | ||
| losses_previous, losses_after, train_accs = update_modes.update_embedding0(wv_twitter_100_emb, n_class, emb_dim, word_list, | ||
| train_pad_encoded, label_train_encoded, test_pad_encoded, | ||
| label_test_encoded) | ||
| epoches1 = [i for i in range(1,16)] | ||
| epoches2 = [i for i in range(16,26)] | ||
|
|
||
| plot2D2(epoches1, losses_previous, epoches2, losses_after, "Epoches", "Accuracy", | ||
| f"Accuracy versus Epoches (Mode - {mode} Update Mode - Setting0)", f"{mode} (before switch) - Setting0", | ||
| f"{mode} (after switch) - Setting1") | ||
| plot2D2(epoches1, losses_previous, epoches2, losses_after, "Epoches", "Accuracy", | ||
| f"Accuracy versus Epoches (Mode - {mode} Update Mode - Setting0)", f"{mode} (before switch) - Setting0", | ||
| f"{mode} (after switch) - Setting0") | ||
| print("UPDATE1 Methods\n") | ||
| if(update_modes!="mt2st_switch"): | ||
| losses, train_accs = update_modes.update_embedding1(wv_twitter_100_emb, n_class, emb_dim, word_list, train_pad_encoded, label_train_encoded, test_pad_encoded, | ||
| label_test_encoded) | ||
| epoches = [i for i in range(1,26)] | ||
| plot2D_loss(epoches, losses, "Epoches", "Loss", | ||
| f"Loss versus Epoches (Execution Mode - {mode} Update Mode - Setting1)", | ||
| f"Execution Mode - {mode} Update Mode - Setting1") | ||
|
|
||
| plot2D_accuracy(epoches, train_accs, "Epoches", "Accuracy", | ||
| f"Accuracy versus Epoches (Execution Mode - {mode} Update Mode - Setting1)", | ||
| f"Execution Mode - {mode} Update Mode - Setting1") | ||
| else: | ||
| losses_previous, losses_after, train_accs = update_modes.update_embedding1(wv_twitter_100_emb, n_class, emb_dim, word_list, | ||
| train_pad_encoded, label_train_encoded, test_pad_encoded, | ||
| label_test_encoded) | ||
|
|
||
| epoches1 = [i for i in range(1,16)] | ||
| epoches2 = [i for i in range(16,26)] | ||
|
|
||
| plot2D2(epoches1, losses_previous, epoches2, losses_after, "Epoches", "Accuracy", | ||
| f"Accuracy versus Epoches (Mode - {mode} Update Mode - Setting1)", f"{mode} (before switch) - Setting1", | ||
| f"{mode} (after switch) - Setting1") | ||
| plot2D2(epoches1, losses_previous, epoches2, losses_after, "Epoches", "Accuracy", | ||
| f"Accuracy versus Epoches (Mode - {mode} Update Mode - Setting1)", f"{mode} (before switch) - Setting1", | ||
| f"{mode} (after switch) - Setting1") | ||
|
|
||
| print("UPDATE2 Methods\n") | ||
| if(update_modes!="mt2st_switch"): | ||
| losses, train_accs = update_modes.update_embedding2(wv_twitter_100_emb, n_class, emb_dim, word_list, train_pad_encoded, label_train_encoded, test_pad_encoded, | ||
| label_test_encoded) | ||
| epoches = [i for i in range(1,26)] | ||
| plot2D_loss(epoches, losses, "Epoches", "Loss", | ||
| f"Loss versus Epoches (Execution Mode - {mode} Update Mode - Setting2)", | ||
| f"Execution Mode - {mode} Update Mode - Setting2") | ||
|
|
||
| plot2D_accuracy(epoches, train_accs, "Epoches", "Accuracy", | ||
| f"Accuracy versus Epoches (Execution Mode - {mode} Update Mode - Setting2)", | ||
| f"Execution Mode - {mode} Update Mode - Setting2") | ||
| else: | ||
| losses_previous, losses_after, train_accs = update_modes.update_embedding2(wv_twitter_100_emb, n_class, emb_dim, word_list, | ||
| train_pad_encoded, label_train_encoded, test_pad_encoded, | ||
| label_test_encoded) | ||
|
|
||
| epoches1 = [i for i in range(1,16)] | ||
| epoches2 = [i for i in range(16,26)] | ||
|
|
||
| plot2D2(epoches1, losses_previous, epoches2, losses_after, "Epoches", "Accuracy", | ||
| f"Accuracy versus Epoches (Mode - {mode} Update Mode - Setting2)", f"{mode} (before switch) - Setting2", | ||
| f"{mode} (after switch) - Setting2") | ||
| plot2D2(epoches1, losses_previous, epoches2, losses_after, "Epoches", "Accuracy", | ||
| f"Accuracy versus Epoches (Mode - {mode} Update Mode - Setting2)", f"{mode} (before switch) - Setting2", | ||
| f"{mode} (after switch) - Setting2") | ||
|
|
||
|
|
||
| print("UPDATE3 Methods\n") | ||
| if(update_modes!="mt2st_switch"): | ||
| losses, train_accs = update_modes.update_embedding3(wv_twitter_100_emb, n_class, emb_dim, word_list, train_pad_encoded, label_train_encoded, test_pad_encoded, | ||
| label_test_encoded) | ||
| epoches = [i for i in range(1,26)] | ||
| plot2D_loss(epoches, losses, "Epoches", "Loss", | ||
| f"Loss versus Epoches (Execution Mode - {mode} Update Mode - Setting3)", | ||
| f"Execution Mode - {mode} Update Mode - Setting3") | ||
|
|
||
| plot2D_accuracy(epoches, train_accs, "Epoches", "Accuracy", | ||
| f"Accuracy versus Epoches (Execution Mode - {mode} Update Mode - Setting3)", | ||
| f"Execution Mode - {mode} Update Mode - Setting3") | ||
| else: | ||
| losses_previous, losses_after, train_accs = update_modes.update_embedding3(wv_twitter_100_emb, n_class, emb_dim, word_list, | ||
| train_pad_encoded, label_train_encoded, test_pad_encoded, | ||
| label_test_encoded) | ||
|
|
||
| epoches1 = [i for i in range(1,16)] | ||
| epoches2 = [i for i in range(16,26)] | ||
| plot2D2(epoches1, losses_previous, epoches2, losses_after, "Epoches", "Accuracy", | ||
| f"Accuracy versus Epoches (Mode - {mode} Update Mode - Setting3)", f"{mode} (before switch) - Setting3", | ||
| f"{mode} (after switch) - Setting3") | ||
| plot2D2(epoches1, losses_previous, epoches2, losses_after, "Epoches", "Accuracy", | ||
| f"Accuracy versus Epoches (Mode - {mode} Update Mode - Setting3)", f"{mode} (before switch) - Setting3", | ||
| f"{mode} (after switch) - Setting3") |
Oops, something went wrong.