Skip to content
Snippets Groups Projects
Commit 4679630e authored by emetheni's avatar emetheni
Browse files

update files

parent 23d2149f
Branches
No related tags found
No related merge requests found
Hide whitespace changes
Inline Side-by-side
classifier_bare_pytorch.py
+ 2
− 2
View file @ 4679630e
......@@ -26,8 +26,8 @@ substitutions_file = 'mappings/substitutions.txt'
# mapping_classes = args.mappings_file[:-4].split('-')[-1]
# specific_results = open_specific_results('mappings/specific_results.txt')['B']
print('\nlangs to use: '+ args.langs_to_use + '\n', flush='True')
print('\nlangs to use: '+ args.langs_to_use)
print('mappings file: ' + args.mappings_file, flush='True')
set_seed(42)
torch.manual_seed(42)
......
classifier_with_adapter.py
+ 42
− 74
View file @ 4679630e
......@@ -29,17 +29,8 @@ mappings, inv_mappings = open_mappings(args.mappings_file)
substitutions_file = 'mappings/substitutions.txt'
tokenizer = AutoTokenizer.from_pretrained(args.transformer_model)
# we are saving the test results of specific epochs
# specific_results = open_specific_results('mappings/specific_results.txt')
# if '1-2-3' in adapter_name or 'layer1;layer2;layer3' in adapter_name:
# specific_results = list(specific_results['A1_3'][args.num_epochs])
# else:
# specific_results = list(specific_results['A1'][args.num_epochs])
set_seed(42)
print('Train classifier with adapter\n')
print('Adapter name:', adapter_name)
print('Model:', args.transformer_model)
......@@ -50,7 +41,12 @@ print('Num epochs:', args.num_epochs)
mappings, inv_mappings = open_mappings(args.mappings_file)
# Open sentences
train_sentences, dev_dict_sentences, test_dict_sentences = open_sentences(args.data_path, mappings)
train_sentences, dev_dict_sentences, test_dict_sentences, framework_labels = open_sentences_with_lang(args.data_path, mappings)
print('\nCheck encodings:\n')
print(train_sentences[0])
# make pandas dataframes
file_header = ['text', 'labels']
......@@ -59,6 +55,13 @@ train_df = pd.DataFrame([[' '.join(x[-2]), x[-1]] for x in train_sentences],
columns =file_header)
train_df = train_df.sample(frac = 1) # shuffle the train
# make a joint dev dataset in order to save models
eval_df = pd.DataFrame([[' '.join(x[-2]), x[-1]]
for corpus, sents in dev_dict_sentences.items()
for x in sents],
columns =file_header)
dev_dict_df = {corpus : pd.DataFrame([[' '.join(x[-2]), x[-1]]
for x in sents],
columns = file_header)
......@@ -71,6 +74,7 @@ test_dict_df = {corpus : pd.DataFrame([[' '.join(x[-2]), x[-1]]
#Make datasets from dataframes
train_dataset = datasets.Dataset.from_pandas(train_df)
eval_dataset = datasets.Dataset.from_pandas(eval_df)
dev_dict_dataset = {corpus:datasets.Dataset.from_pandas(dev_df)
for corpus, dev_df in dev_dict_df.items()}
test_dict_dataset = {corpus:datasets.Dataset.from_pandas(dev_df)
......@@ -84,6 +88,9 @@ num_labels = len(set([int(x.strip())
train_dataset = train_dataset.map(encode_batch, batched=True)
train_dataset.set_format(type="torch", columns=["input_ids", "attention_mask", "labels"])
eval_dataset = eval_dataset.map(encode_batch, batched=True)
eval_dataset.set_format(type="torch", columns=["input_ids", "attention_mask", "labels"])
encoded_dev_dataset = {}
for corpus in dev_dict_dataset:
temp = dev_dict_dataset[corpus].map(encode_batch, batched=True)
......@@ -129,7 +136,8 @@ training_args = TrainingArguments(
trainer = Trainer(
model = model,
args = training_args,
train_dataset = train_dataset
train_dataset = train_dataset,
eval_dataset = eval_dataset
)
......@@ -153,72 +161,32 @@ trainer.train()
print('\nDev results:')
for corpus in encoded_dev_dataset:
print()
dev_results = get_predictions_huggingface(trainer, corpus,
encoded_dev_dataset[corpus])
dev_results_ = get_predictions_huggingface(trainer,
corpus,
encoded_dev_dataset[corpus]
)
path_results = 'results/dev/' + adapter_name + '_' + str(args.num_epochs)
if not os.path.exists(path_results):
os.makedirs(path_results)
print_results_to_file(corpus,
dev_dict_sentences[corpus],
dev_results,
inv_mappings,
#substitutions_file,
path_results)
# Test results
print('\ntest results:')
dev_results = better_predictions_huggingface(trainer,
corpus,
encoded_dev_dataset[corpus],
framework_labels[corpus.split('.')[1]]
)
print('\nTest results:')
for corpus in encoded_test_dataset:
print()
test_results = get_predictions_huggingface(trainer,
corpus,
encoded_test_dataset[corpus])
path_results = 'results/test/' + adapter_name + '_' + str(args.num_epochs)
if not os.path.exists(path_results):
os.makedirs(path_results)
print_results_to_file(corpus,
test_dict_sentences[corpus],
test_results,
inv_mappings,
#substitutions_file,
path_results)
# for corpus in test_dict_dataloader:
# test_results = get_predictions(model,
# corpus,
# test_dict_dataloader[corpus])
# path_results = 'results/test/pytorch' + str(epoch_num+1)
# if not os.path.exists(path_results):
# os.makedirs(path_results)
# print_results_to_file(corpus,
# test_dict_sentences[corpus],
# test_results,
# inv_mappings, substitutions_file,
# path_results)
dev_results_ = get_predictions_huggingface(trainer,
corpus,
encoded_test_dataset[corpus]
)
# Save specific test results
# print('\nTest results:')
# for corpus in encoded_test_dataset:
# print()
# test_results = get_predictions_huggingface(trainer, corpus,
# encoded_test_dataset[corpus])
#
# print_results_to_file(corpus, test_dict_sentences[corpus], test_results,
# inv_mappings, substitutions_file)
\ No newline at end of file
dev_results = better_predictions_huggingface(trainer,
corpus,
encoded_test_dataset[corpus],
framework_labels[corpus.split('.')[1]]
)
utils.py
+ 10
− 4
View file @ 4679630e
......@@ -6,7 +6,8 @@ import torch
from transformers import AutoConfig, AutoTokenizer
from configure import parse_args
import numpy as np
from sklearn.metrics import accuracy_score
from sklearn.metrics import accuracy_score, confusion_matrix, classification_report
import matplotlib.pyplot as plt
args = parse_args()
......@@ -333,6 +334,14 @@ def open_sentences_with_lang(path_to_corpora, mappings_dict):
return train_sentences, dev_dict_sentences, test_dict_sentences, corpus_labels
def show_confusion(predictions, labels, save_name):
cm = confusion_matrix(y_test, y_pred)
cm_display = ConfusionMatrixDisplay(cm).plot()
plt.savefig(save_name, dpi=300)
return None
# ===============
# Testing functions
......@@ -417,9 +426,6 @@ def better_predictions_huggingface(trainer,
results = trainer.predict(test_set)
orig_labels = results.label_ids.tolist()
print('len sentences', len(orig_labels))
print('shape of preds', results.predictions.shape)
results_per_sent = results.predictions.tolist()
# try to make the better prediction bit
......
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Please register or to comment