055: Adding word level examples

21 files changed, 93 insertions, 311 deletions

diff --git a/examples/app/__init__.py b/examples/app/__init__.py
deleted file mode 100644
index e69de29..0000000
--- a/examples/app/__init__.py
+++ /dev/null
diff --git a/examples/app/monotransquest_app_test.py b/examples/app/monotransquest_app_test.py
deleted file mode 100644
index 433bfce..0000000
--- a/examples/app/monotransquest_app_test.py
+++ /dev/null
@@ -1,17 +0,0 @@
-import logging
-
-from transquest.app.monotransquest_app import MonoTransQuestApp
-
-
-logging.basicConfig()
-logging.getLogger().setLevel(logging.INFO)
-
-test_sentences = [
-    [
-        "Jocurile de oferă noi provocări pentru IA în domeniul teoriei jocurilor.",
-        "Games provide new challenges for IA in the area of gambling theory"
-    ]
-]
-
-app = MonoTransQuestApp("monotransquest-da-ro_en", use_cuda=False, force_download=True)
-print(app.predict_quality(test_sentences))
diff --git a/transquest/algo/sentence_level/monotransquest/models/bert_model.py b/transquest/algo/sentence_level/monotransquest/models/bert_model.py
index ca20dd6..b8ec4a4 100755
--- a/transquest/algo/sentence_level/monotransquest/models/bert_model.py
+++ b/transquest/algo/sentence_level/monotransquest/models/bert_model.py
@@ -1,4 +1,3 @@
-import torch
 import torch.nn as nn
 from torch.nn import CrossEntropyLoss, MSELoss
 from transformers.models.bert.modeling_bert import BertModel, BertPreTrainedModel
@@ -44,14 +43,14 @@ class BertForSequenceClassification(BertPreTrainedModel):
         self.init_weights()
 
     def forward(
-        self,
-        input_ids=None,
-        attention_mask=None,
-        token_type_ids=None,
-        position_ids=None,
-        head_mask=None,
-        inputs_embeds=None,
-        labels=None,
+            self,
+            input_ids=None,
+            attention_mask=None,
+            token_type_ids=None,
+            position_ids=None,
+            head_mask=None,
+            inputs_embeds=None,
+            labels=None,
     ):
 
         outputs = self.bert(
diff --git a/transquest/algo/sentence_level/monotransquest/models/distilbert_model.py b/transquest/algo/sentence_level/monotransquest/models/distilbert_model.py
index 4de0b38..2710618 100755
--- a/transquest/algo/sentence_level/monotransquest/models/distilbert_model.py
+++ b/transquest/algo/sentence_level/monotransquest/models/distilbert_model.py
@@ -1,5 +1,5 @@
 import torch.nn as nn
-from torch.nn import CrossEntropyLoss, MSELoss
+from torch.nn import CrossEntropyLoss
 from transformers.models.distilbert.modeling_distilbert import DistilBertModel, DistilBertPreTrainedModel
 
 
@@ -44,7 +44,8 @@ class DistilBertForSequenceClassification(DistilBertPreTrainedModel):
         self.init_weights()
 
     def forward(
-        self, input_ids=None, attention_mask=None, head_mask=None, inputs_embeds=None, labels=None, class_weights=None,
+            self, input_ids=None, attention_mask=None, head_mask=None, inputs_embeds=None, labels=None,
+            class_weights=None,
     ):
         distilbert_output = self.distilbert(input_ids=input_ids, attention_mask=attention_mask, head_mask=head_mask)
         hidden_state = distilbert_output[0]  # (bs, seq_len, dim)
diff --git a/transquest/algo/sentence_level/monotransquest/models/roberta_model.py b/transquest/algo/sentence_level/monotransquest/models/roberta_model.py
index 6d10790..41b190a 100755
--- a/transquest/algo/sentence_level/monotransquest/models/roberta_model.py
+++ b/transquest/algo/sentence_level/monotransquest/models/roberta_model.py
@@ -1,6 +1,5 @@
-import torch
-import torch.nn as nn
 from torch.nn import CrossEntropyLoss, MSELoss
+from transformers import BertPreTrainedModel
 from transformers.models.roberta.modeling_roberta import (
     ROBERTA_PRETRAINED_MODEL_ARCHIVE_LIST,
     RobertaClassificationHead,
@@ -8,8 +7,6 @@ from transformers.models.roberta.modeling_roberta import (
     RobertaModel,
 )
 
-from transformers import BertPreTrainedModel
-
 
 class RobertaForSequenceClassification(BertPreTrainedModel):
     r"""
@@ -51,14 +48,14 @@ class RobertaForSequenceClassification(BertPreTrainedModel):
         self.weight = weight
 
     def forward(
-        self,
-        input_ids=None,
-        attention_mask=None,
-        token_type_ids=None,
-        position_ids=None,
-        head_mask=None,
-        inputs_embeds=None,
-        labels=None,
+            self,
+            input_ids=None,
+            attention_mask=None,
+            token_type_ids=None,
+            position_ids=None,
+            head_mask=None,
+            inputs_embeds=None,
+            labels=None,
     ):
         outputs = self.roberta(
             input_ids,
diff --git a/transquest/algo/sentence_level/monotransquest/models/xlm_model.py b/transquest/algo/sentence_level/monotransquest/models/xlm_model.py
index 6bb9fca..c7c8380 100755
--- a/transquest/algo/sentence_level/monotransquest/models/xlm_model.py
+++ b/transquest/algo/sentence_level/monotransquest/models/xlm_model.py
@@ -1,5 +1,3 @@
-import torch
-import torch.nn as nn
 from torch.nn import CrossEntropyLoss, MSELoss
 from transformers.models.xlm.modeling_xlm import SequenceSummary, XLMModel, XLMPreTrainedModel
 
@@ -43,17 +41,17 @@ class XLMForSequenceClassification(XLMPreTrainedModel):
         self.init_weights()
 
     def forward(
-        self,
-        input_ids=None,
-        attention_mask=None,
-        langs=None,
-        token_type_ids=None,
-        position_ids=None,
-        lengths=None,
-        cache=None,
-        head_mask=None,
-        inputs_embeds=None,
-        labels=None,
+            self,
+            input_ids=None,
+            attention_mask=None,
+            langs=None,
+            token_type_ids=None,
+            position_ids=None,
+            lengths=None,
+            cache=None,
+            head_mask=None,
+            inputs_embeds=None,
+            labels=None,
     ):
         transformer_outputs = self.transformer(
             input_ids,
diff --git a/transquest/algo/sentence_level/siamesetransquest/datasets/sentence_label_dataset.py b/transquest/algo/sentence_level/siamesetransquest/datasets/sentence_label_dataset.py
index 09cb6aa..66b7ba7 100644
--- a/transquest/algo/sentence_level/siamesetransquest/datasets/sentence_label_dataset.py
+++ b/transquest/algo/sentence_level/siamesetransquest/datasets/sentence_label_dataset.py
@@ -7,8 +7,8 @@ import torch
 from torch.utils.data import Dataset
 from tqdm import tqdm
 
-from transquest.algo.sentence_level.siamesetransquest.run_model import SiameseTransQuestModel
 from transquest.algo.sentence_level.siamesetransquest.readers.input_example import InputExample
+from transquest.algo.sentence_level.siamesetransquest.run_model import SiameseTransQuestModel
 
 
 class SentenceLabelDataset(Dataset):
@@ -84,7 +84,7 @@ class SentenceLabelDataset(Dataset):
 
             if hasattr(model,
                        'max_seq_length') and model.max_seq_length is not None and model.max_seq_length > 0 and len(
-                    tokenized_text) >= model.max_seq_length:
+                tokenized_text) >= model.max_seq_length:
                 too_long += 1
             if example.label in label_sent_mapping:
                 label_sent_mapping[example.label].append(ex_index)
diff --git a/transquest/algo/sentence_level/siamesetransquest/datasets/sentences_dataset.py b/transquest/algo/sentence_level/siamesetransquest/datasets/sentences_dataset.py
index 4c4b693..d964ea5 100644
--- a/transquest/algo/sentence_level/siamesetransquest/datasets/sentences_dataset.py
+++ b/transquest/algo/sentence_level/siamesetransquest/datasets/sentences_dataset.py
@@ -5,8 +5,8 @@ import torch
 from torch.utils.data import Dataset
 from tqdm import tqdm
 
-from transquest.algo.sentence_level.siamesetransquest.run_model import SiameseTransQuestModel
 from transquest.algo.sentence_level.siamesetransquest.readers.input_example import InputExample
+from transquest.algo.sentence_level.siamesetransquest.run_model import SiameseTransQuestModel
 
 
 class SentencesDataset(Dataset):
@@ -23,7 +23,7 @@ class SentencesDataset(Dataset):
         """
         if show_progress_bar is None:
             show_progress_bar = (
-                        logging.getLogger().getEffectiveLevel() == logging.INFO or logging.getLogger().getEffectiveLevel() == logging.DEBUG)
+                    logging.getLogger().getEffectiveLevel() == logging.INFO or logging.getLogger().getEffectiveLevel() == logging.DEBUG)
         self.show_progress_bar = show_progress_bar
 
         self.convert_input_examples(examples, model)
diff --git a/transquest/algo/sentence_level/siamesetransquest/evaluation/binary_embedding_similarity_evaluator.py b/transquest/algo/sentence_level/siamesetransquest/evaluation/binary_embedding_similarity_evaluator.py
index cb773fe..770f1e0 100644
--- a/transquest/algo/sentence_level/siamesetransquest/evaluation/binary_embedding_similarity_evaluator.py
+++ b/transquest/algo/sentence_level/siamesetransquest/evaluation/binary_embedding_similarity_evaluator.py
@@ -8,8 +8,8 @@ from sklearn.metrics.pairwise import paired_cosine_distances, paired_euclidean_d
 from torch.utils.data import DataLoader
 from tqdm import tqdm
 
-from transquest.algo.sentence_level.siamesetransquest.evaluation import SimilarityFunction
 from transquest.algo.sentence_level.siamesetransquest.evaluation import SentenceEvaluator
+from transquest.algo.sentence_level.siamesetransquest.evaluation import SimilarityFunction
 from transquest.algo.sentence_level.siamesetransquest.util import batch_to_device
 
 
diff --git a/transquest/algo/sentence_level/siamesetransquest/evaluation/embedding_similarity_evaluator.py b/transquest/algo/sentence_level/siamesetransquest/evaluation/embedding_similarity_evaluator.py
index 8d6d7d6..b63a95b 100644
--- a/transquest/algo/sentence_level/siamesetransquest/evaluation/embedding_similarity_evaluator.py
+++ b/transquest/algo/sentence_level/siamesetransquest/evaluation/embedding_similarity_evaluator.py
@@ -9,8 +9,8 @@ from sklearn.metrics.pairwise import paired_cosine_distances, paired_euclidean_d
 from torch.utils.data import DataLoader
 from tqdm import tqdm
 
-from transquest.algo.sentence_level.siamesetransquest.evaluation import SimilarityFunction
 from transquest.algo.sentence_level.siamesetransquest.evaluation import SentenceEvaluator
+from transquest.algo.sentence_level.siamesetransquest.evaluation import SimilarityFunction
 from transquest.algo.sentence_level.siamesetransquest.util import batch_to_device
 
 
@@ -44,7 +44,7 @@ class EmbeddingSimilarityEvaluator(SentenceEvaluator):
 
         if show_progress_bar is None:
             show_progress_bar = (
-                        logging.getLogger().getEffectiveLevel() == logging.INFO or logging.getLogger().getEffectiveLevel() == logging.DEBUG)
+                    logging.getLogger().getEffectiveLevel() == logging.INFO or logging.getLogger().getEffectiveLevel() == logging.DEBUG)
         self.show_progress_bar = show_progress_bar
 
         self.device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
diff --git a/transquest/algo/sentence_level/siamesetransquest/evaluation/translation_evaluator.py b/transquest/algo/sentence_level/siamesetransquest/evaluation/translation_evaluator.py
index 36382f5..09f2b49 100644
--- a/transquest/algo/sentence_level/siamesetransquest/evaluation/translation_evaluator.py
+++ b/transquest/algo/sentence_level/siamesetransquest/evaluation/translation_evaluator.py
@@ -8,8 +8,8 @@ import torch
 from torch.utils.data import DataLoader
 from tqdm import tqdm
 
-from transquest.algo.sentence_level.siamesetransquest.evaluation import SimilarityFunction
 from transquest.algo.sentence_level.siamesetransquest.evaluation import SentenceEvaluator
+from transquest.algo.sentence_level.siamesetransquest.evaluation import SimilarityFunction
 from transquest.algo.sentence_level.siamesetransquest.util import batch_to_device
 
 
@@ -40,7 +40,7 @@ class TranslationEvaluator(SentenceEvaluator):
 
         if show_progress_bar is None:
             show_progress_bar = (
-                        logging.getLogger().getEffectiveLevel() == logging.INFO or logging.getLogger().getEffectiveLevel() == logging.DEBUG)
+                    logging.getLogger().getEffectiveLevel() == logging.INFO or logging.getLogger().getEffectiveLevel() == logging.DEBUG)
         self.show_progress_bar = show_progress_bar
 
         self.device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
diff --git a/transquest/algo/sentence_level/siamesetransquest/evaluation/triplet_evaluator.py b/transquest/algo/sentence_level/siamesetransquest/evaluation/triplet_evaluator.py
index 2be2c69..95590f1 100644
--- a/transquest/algo/sentence_level/siamesetransquest/evaluation/triplet_evaluator.py
+++ b/transquest/algo/sentence_level/siamesetransquest/evaluation/triplet_evaluator.py
@@ -7,8 +7,8 @@ from sklearn.metrics.pairwise import paired_cosine_distances, paired_euclidean_d
 from torch.utils.data import DataLoader
 from tqdm import tqdm
 
-from transquest.algo.sentence_level.siamesetransquest.evaluation import SimilarityFunction
 from transquest.algo.sentence_level.siamesetransquest.evaluation import SentenceEvaluator
+from transquest.algo.sentence_level.siamesetransquest.evaluation import SimilarityFunction
 
 
 class TripletEvaluator(SentenceEvaluator):
diff --git a/transquest/algo/sentence_level/siamesetransquest/run_model.py b/transquest/algo/sentence_level/siamesetransquest/run_model.py
index d9bac73..d485465 100644
--- a/transquest/algo/sentence_level/siamesetransquest/run_model.py
+++ b/transquest/algo/sentence_level/siamesetransquest/run_model.py
@@ -116,7 +116,7 @@ class SiameseTransQuestModel(nn.Sequential):
         self.eval()
         if show_progress_bar is None:
             show_progress_bar = (
-                        logging.getLogger().getEffectiveLevel() == logging.INFO or logging.getLogger().getEffectiveLevel() == logging.DEBUG)
+                    logging.getLogger().getEffectiveLevel() == logging.INFO or logging.getLogger().getEffectiveLevel() == logging.DEBUG)
 
         all_embeddings = []
         length_sorted_idx = np.argsort([len(sen) for sen in sentences])
diff --git a/transquest/algo/word_level/microtransquest/format.py b/transquest/algo/word_level/microtransquest/format.py
index 1543dd6..607c2c6 100644
--- a/transquest/algo/word_level/microtransquest/format.py
+++ b/transquest/algo/word_level/microtransquest/format.py
@@ -9,7 +9,8 @@ def prepare_data(raw_df, args):
 
     sentence_id = 0
     data = []
-    for source_sentence, source_tag_line, target_sentence, target_tag_lind in zip(source_sentences, source_tags, target_sentences, target_tags):
+    for source_sentence, source_tag_line, target_sentence, target_tag_lind in zip(source_sentences, source_tags,
+                                                                                  target_sentences, target_tags):
         for word, tag in zip(source_sentence.split(), source_tag_line.split()):
             data.append([sentence_id, word, tag])
 
@@ -18,7 +19,6 @@ def prepare_data(raw_df, args):
         target_words = target_sentence.split()
         target_tags = target_tag_lind.split()
 
-
         data.append([sentence_id, args["tag"], target_tags.pop(0)])
 
         for word in target_words:
@@ -31,7 +31,6 @@ def prepare_data(raw_df, args):
 
 
 def prepare_testdata(raw_df, args):
-
     source_sentences = raw_df[args["source_column"]].tolist()
     target_sentences = raw_df[args["target_column"]].tolist()
 
@@ -77,7 +76,8 @@ def post_process(predicted_sentences, test_sentences, args):
         assert len(source_tags) == len(source_sentence.split())
 
         if len(target_sentence.split()) > len(target_tags):
-            target_tags = target_tags + [args["default_quality"] for x in range(len(target_sentence.split()) - len(target_tags))]
+            target_tags = target_tags + [args["default_quality"] for x in
+                                         range(len(target_sentence.split()) - len(target_tags))]
 
         assert len(target_tags) == len(target_sentence.split())
         sources_tags.append(source_tags)
@@ -85,7 +85,6 @@ def post_process(predicted_sentences, test_sentences, args):
 
     return sources_tags, targets_tags
 
-
 # def post_process(predicted_sentences, test_sentences):
 #     sources_tags = []
 #     targets_tags = []
@@ -108,7 +107,3 @@ def post_process(predicted_sentences, test_sentences, args):
 #         targets_tags.append(target_tags)
 #
 #     return sources_tags, targets_tags
-
-
-
-
diff --git a/transquest/algo/word_level/microtransquest/model_args.py b/transquest/algo/word_level/microtransquest/model_args.py
index 3f81826..4a5ca66 100644
--- a/transquest/algo/word_level/microtransquest/model_args.py
+++ b/transquest/algo/word_level/microtransquest/model_args.py
@@ -19,4 +19,3 @@ class MicroTransQuestArgs(TransQuestArgs):
     add_tag: bool = False
     tag: str = "<gap>"
     default_quality: str = "OK"
-
diff --git a/transquest/algo/word_level/microtransquest/run_model.py b/transquest/algo/word_level/microtransquest/run_model.py
index 57c5612..d75b184 100755
--- a/transquest/algo/word_level/microtransquest/run_model.py
+++ b/transquest/algo/word_level/microtransquest/run_model.py
@@ -1,72 +1,48 @@
 from __future__ import absolute_import, division, print_function
 
 import glob
-import json
 import logging
 import math
 import os
 import random
 import shutil
+import tempfile
 import warnings
 from dataclasses import asdict
-from multiprocessing import cpu_count
-import tempfile
 from pathlib import Path
 
 import numpy as np
 import pandas as pd
 import torch
-from scipy.stats import pearsonr
 from seqeval.metrics import classification_report, f1_score, precision_score, recall_score
 from tensorboardX import SummaryWriter
 from torch.nn import CrossEntropyLoss
 from torch.utils.data import DataLoader, RandomSampler, SequentialSampler, TensorDataset
 from tqdm.auto import tqdm, trange
-from transformers.optimization import (
-    get_constant_schedule,
-    get_constant_schedule_with_warmup,
-    get_linear_schedule_with_warmup,
-    get_cosine_schedule_with_warmup,
-    get_cosine_with_hard_restarts_schedule_with_warmup,
-    get_polynomial_decay_schedule_with_warmup,
-)
-from transformers.optimization import AdamW, Adafactor
 from transformers import (
-    WEIGHTS_NAME,
-    AutoConfig,
-    AutoModelForTokenClassification,
-    AutoTokenizer,
     BertConfig,
     BertForTokenClassification,
     BertTokenizer,
-    BertweetTokenizer,
-    CamembertConfig,
-    CamembertForTokenClassification,
-    CamembertTokenizer,
     DistilBertConfig,
     DistilBertForTokenClassification,
     DistilBertTokenizer,
-    ElectraConfig,
-    ElectraForTokenClassification,
-    ElectraTokenizer,
-    LongformerConfig,
-    LongformerForTokenClassification,
-    LongformerTokenizer,
-    MobileBertConfig,
-    MobileBertForTokenClassification,
-    MobileBertTokenizer,
     RobertaConfig,
     RobertaForTokenClassification,
     RobertaTokenizer,
     XLMRobertaConfig,
     XLMRobertaForTokenClassification,
     XLMRobertaTokenizer,
-    LayoutLMConfig,
-    LayoutLMForTokenClassification,
-    LayoutLMTokenizer,
 )
-from wandb import config
 from transformers.convert_graph_to_onnx import convert, quantize
+from transformers.optimization import AdamW, Adafactor
+from transformers.optimization import (
+    get_constant_schedule,
+    get_constant_schedule_with_warmup,
+    get_linear_schedule_with_warmup,
+    get_cosine_schedule_with_warmup,
+    get_cosine_with_hard_restarts_schedule_with_warmup,
+    get_polynomial_decay_schedule_with_warmup,
+)
 
 from transquest.algo.word_level.microtransquest.model_args import MicroTransQuestArgs
 from transquest.algo.word_level.microtransquest.utils import sweep_config_to_sweep_values, InputExample, \
@@ -84,15 +60,15 @@ logger = logging.getLogger(__name__)
 
 class MicroTransQuestModel:
     def __init__(
-        self,
-        model_type,
-        model_name,
-        labels=None,
-        args=None,
-        use_cuda=True,
-        cuda_device=-1,
-        onnx_execution_provider=None,
-        **kwargs,
+            self,
+            model_type,
+            model_name,
+            labels=None,
+            args=None,
+            use_cuda=True,
+            cuda_device=-1,
+            onnx_execution_provider=None,
+            **kwargs,
     ):
         """
         Initializes a NERModel
@@ -251,7 +227,7 @@ class MicroTransQuestModel:
             self.args.wandb_project = None
 
     def train_model(
-        self, train_data, output_dir=None, show_running_loss=True, args=None, eval_data=None, verbose=True, **kwargs
+            self, train_data, output_dir=None, show_running_loss=True, args=None, eval_data=None, verbose=True, **kwargs
     ):
         """
         Trains the model using 'train_data'
@@ -483,7 +459,7 @@ class MicroTransQuestModel:
                 global_step = int(checkpoint_suffix)
                 epochs_trained = global_step // (len(train_dataloader) // args.gradient_accumulation_steps)
                 steps_trained_in_current_epoch = global_step % (
-                    len(train_dataloader) // args.gradient_accumulation_steps
+                        len(train_dataloader) // args.gradient_accumulation_steps
                 )
 
                 logger.info("   Continuing training from checkpoint, will skip to saved global_step")
@@ -596,8 +572,8 @@ class MicroTransQuestModel:
                         self.save_model(output_dir_current, optimizer, scheduler, model=model)
 
                     if args.evaluate_during_training and (
-                        args.evaluate_during_training_steps > 0
-                        and global_step % args.evaluate_during_training_steps == 0
+                            args.evaluate_during_training_steps > 0
+                            and global_step % args.evaluate_during_training_steps == 0
                     ):
 
                         output_dir_current = os.path.join(output_dir, "checkpoint-{}".format(global_step))
@@ -1031,7 +1007,7 @@ class MicroTransQuestModel:
                 else:
                     preds = np.append(preds, output[0], axis=0)
                     out_input_ids = np.append(out_input_ids, inputs_onnx["input_ids"], axis=0)
-                    out_attention_mask = np.append(out_attention_mask, inputs_onnx["attention_mask"], axis=0,)
+                    out_attention_mask = np.append(out_attention_mask, inputs_onnx["attention_mask"], axis=0, )
             out_label_ids = np.zeros_like(out_input_ids)
             for index in range(len(out_label_ids)):
                 out_label_ids[index][0] = -100
@@ -1209,8 +1185,8 @@ class MicroTransQuestModel:
                 os.makedirs(self.args.cache_dir, exist_ok=True)
 
             if os.path.exists(cached_features_file) and (
-                (not args.reprocess_input_data and not no_cache)
-                or (mode == "dev" and args.use_cached_eval_features and not no_cache)
+                    (not args.reprocess_input_data and not no_cache)
+                    or (mode == "dev" and args.use_cached_eval_features and not no_cache)
             ):
                 features = torch.load(cached_features_file)
                 logger.info(f" Features loaded from cache at {cached_features_file}")
diff --git a/transquest/algo/word_level/microtransquest/utils.py b/transquest/algo/word_level/microtransquest/utils.py
index 774079b..372a18a 100755
--- a/transquest/algo/word_level/microtransquest/utils.py
+++ b/transquest/algo/word_level/microtransquest/utils.py
@@ -18,12 +18,9 @@
 from __future__ import absolute_import, division, print_function
 
 import linecache
-import logging
-import os
 from io import open
 from multiprocessing import Pool, cpu_count
 
-import pandas as pd
 import torch
 from torch.functional import split
 from torch.nn import CrossEntropyLoss
@@ -161,7 +158,7 @@ def get_examples_from_df(data, bbox=False):
         ]
     else:
         return [
-            InputExample(guid=sentence_id, words=sentence_df["words"].tolist(), labels=sentence_df["labels"].tolist(),)
+            InputExample(guid=sentence_id, words=sentence_df["words"].tolist(), labels=sentence_df["labels"].tolist(), )
             for sentence_id, sentence_df in data.groupby(["sentence_id"])
         ]
 
@@ -290,29 +287,29 @@ def convert_example_to_feature(example_row):
             input_ids=input_ids, input_mask=input_mask, segment_ids=segment_ids, label_ids=label_ids, bboxes=bboxes
         )
     else:
-        return InputFeatures(input_ids=input_ids, input_mask=input_mask, segment_ids=segment_ids, label_ids=label_ids,)
+        return InputFeatures(input_ids=input_ids, input_mask=input_mask, segment_ids=segment_ids, label_ids=label_ids, )
 
 
 def convert_examples_to_features(
-    examples,
-    label_list,
-    max_seq_length,
-    tokenizer,
-    cls_token_at_end=False,
-    cls_token="[CLS]",
-    cls_token_segment_id=1,
-    sep_token="[SEP]",
-    sep_token_extra=False,
-    pad_on_left=False,
-    pad_token=0,
-    pad_token_segment_id=0,
-    pad_token_label_id=-1,
-    sequence_a_segment_id=0,
-    mask_padding_with_zero=True,
-    process_count=cpu_count() - 2,
-    chunksize=500,
-    silent=False,
-    use_multiprocessing=True,
+        examples,
+        label_list,
+        max_seq_length,
+        tokenizer,
+        cls_token_at_end=False,
+        cls_token="[CLS]",
+        cls_token_segment_id=1,
+        sep_token="[SEP]",
+        sep_token_extra=False,
+        pad_on_left=False,
+        pad_token=0,
+        pad_token_segment_id=0,
+        pad_token_label_id=-1,
+        sequence_a_segment_id=0,
+        mask_padding_with_zero=True,
+        process_count=cpu_count() - 2,
+        chunksize=500,
+        silent=False,
+        use_multiprocessing=True,
 ):
     """ Loads a data file into a list of `InputBatch`s
         `cls_token_at_end` define the location of the CLS token:
diff --git a/transquest/app/__init__.py b/transquest/app/__init__.py
deleted file mode 100644
index e69de29..0000000
--- a/transquest/app/__init__.py
+++ /dev/null
diff --git a/transquest/app/monotransquest_app.py b/transquest/app/monotransquest_app.py
deleted file mode 100644
index 8187259..0000000
--- a/transquest/app/monotransquest_app.py
+++ /dev/null
@@ -1,59 +0,0 @@
-import logging
-import os
-
-from transquest.algo.sentence_level.monotransquest.run_model import MonoTransQuestModel
-from transquest.app.util.model_downloader import GoogleDriveDownloader as gdd
-
-logger = logging.getLogger(__name__)
-
-
-class MonoTransQuestApp:
-    def __init__(self, model_name_or_path, model_type=None, use_cuda=True, force_download=False, cuda_device=-1):
-
-        self.model_name_or_path = model_name_or_path
-        self.model_type = model_type
-        self.use_cuda = use_cuda
-        self.cuda_device = cuda_device
-
-
-        MODEL_CONFIG = {
-            "monotransquest-da-si_en": ("xlmroberta", "1-UXvna_RGnb6_TTRr4vSGCqA5yl0SYn9", 3.8),
-            "monotransquest-da-ro_en": ("xlmroberta", "1-aeDbR_ftqsTslFJbNybebj5MAhPfIw8", 3.8)
-        }
-
-        if model_name_or_path in MODEL_CONFIG:
-            self.trained_model_type, self.drive_id, self.size = MODEL_CONFIG[model_name_or_path]
-
-            try:
-                from torch.hub import _get_torch_home
-                torch_cache_home = _get_torch_home()
-            except ImportError:
-                torch_cache_home = os.path.expanduser(
-                    os.getenv('TORCH_HOME', os.path.join(
-                        os.getenv('XDG_CACHE_HOME', '~/.cache'), 'torch')))
-            default_cache_path = os.path.join(torch_cache_home, 'transquest')
-            self.model_path = os.path.join(default_cache_path, self.model_name_or_path)
-            if force_download or (not os.path.exists(self.model_path) or not os.listdir(self.model_path)):
-                logger.info(
-                    "Downloading a MonoTransQuest model and saving it at {}".format(self.model_path))
-
-                gdd.download_file_from_google_drive(file_id=self.drive_id,
-                                                    dest_path=os.path.join(self.model_path, "model.zip"),
-                                                    showsize=True, unzip=True, overwrite=True)
-
-            self.model = MonoTransQuestModel(self.trained_model_type, self.model_path, use_cuda=self.use_cuda,
-                                             cuda_device=self.cuda_device)
-
-        else:
-            self.model = MonoTransQuestModel(model_type, self.model_name_or_path, use_cuda=self.use_cuda,
-                                             cuda_device=self.cuda_device)
-
-    @staticmethod
-    def _download(drive_id, model_name):
-        gdd.download_file_from_google_drive(file_id=drive_id,
-                                            dest_path=os.path.join(".transquest", model_name, "model.zip"),
-                                            unzip=True)
-
-    def predict_quality(self, test_sentence_pairs):
-        predictions, raw_outputs = self.model.predict(test_sentence_pairs)
-        return predictions
diff --git a/transquest/app/util/__init__.py b/transquest/app/util/__init__.py
deleted file mode 100644
index e69de29..0000000
--- a/transquest/app/util/__init__.py
+++ /dev/null
diff --git a/transquest/app/util/model_downloader.py b/transquest/app/util/model_downloader.py
deleted file mode 100644
index 9999db2..0000000
--- a/transquest/app/util/model_downloader.py
+++ /dev/null
@@ -1,104 +0,0 @@
-from __future__ import print_function
-import requests
-import zipfile
-import warnings
-from sys import stdout
-from os import makedirs
-from os.path import dirname
-from os.path import exists
-
-
-class GoogleDriveDownloader:
-    """
-    Minimal class to download shared files from Google Drive.
-    """
-
-    CHUNK_SIZE = 32768
-    DOWNLOAD_URL = 'https://docs.google.com/uc?export=download'
-
-    @staticmethod
-    def download_file_from_google_drive(file_id, dest_path, overwrite=False, unzip=False, showsize=False):
-        """
-        Downloads a shared file from google drive into a given folder.
-        Optionally unzips it.
-        Parameters
-        ----------
-        file_id: str
-            the file identifier.
-            You can obtain it from the sharable link.
-        dest_path: str
-            the destination where to save the downloaded file.
-            Must be a path (for example: './downloaded_file.txt')
-        overwrite: bool
-            optional, if True forces re-download and overwrite.
-        unzip: bool
-            optional, if True unzips a file.
-            If the file is not a zip file, ignores it.
-        showsize: bool
-            optional, if True print the current download size.
-        Returns
-        -------
-        None
-        """
-
-        destination_directory = dirname(dest_path)
-        if not exists(destination_directory):
-            makedirs(destination_directory)
-
-        if not exists(dest_path) or overwrite:
-
-            session = requests.Session()
-
-            print('Downloading {} into {}... '.format(file_id, dest_path), end='')
-            stdout.flush()
-
-            response = session.get(GoogleDriveDownloader.DOWNLOAD_URL, params={'id': file_id}, stream=True)
-
-            token = GoogleDriveDownloader._get_confirm_token(response)
-            if token:
-                params = {'id': file_id, 'confirm': token}
-                response = session.get(GoogleDriveDownloader.DOWNLOAD_URL, params=params, stream=True)
-
-            if showsize:
-                print()  # Skip to the next line
-
-            current_download_size = [0]
-            GoogleDriveDownloader._save_response_content(response, dest_path, showsize, current_download_size)
-            print('Done.')
-
-            if unzip:
-                try:
-                    print('Unzipping...', end='')
-                    stdout.flush()
-                    with zipfile.ZipFile(dest_path, 'r') as z:
-                        z.extractall(destination_directory)
-                    print('Done.')
-                except zipfile.BadZipfile:
-                    warnings.warn('Ignoring `unzip` since "{}" does not look like a valid zip file'.format(file_id))
-
-    @staticmethod
-    def _get_confirm_token(response):
-        for key, value in response.cookies.items():
-            if key.startswith('download_warning'):
-                return value
-        return None
-
-    @staticmethod
-    def _save_response_content(response, destination, showsize, current_size):
-        with open(destination, 'wb') as f:
-            for chunk in response.iter_content(GoogleDriveDownloader.CHUNK_SIZE):
-                if chunk:  # filter out keep-alive new chunks
-                    f.write(chunk)
-                    if showsize:
-                        print('\r' + GoogleDriveDownloader.sizeof_fmt(current_size[0]), end=' ')
-                        stdout.flush()
-                        current_size[0] += GoogleDriveDownloader.CHUNK_SIZE
-
-    # From https://stackoverflow.com/questions/1094841/reusable-library-to-get-human-readable-version-of-file-size
-    @staticmethod
-    def sizeof_fmt(num, suffix='B'):
-        for unit in ['', 'Ki', 'Mi', 'Gi', 'Ti', 'Pi', 'Ei', 'Zi']:
-            if abs(num) < 1024.0:
-                return '{:.1f} {}{}'.format(num, unit, suffix)
-            num /= 1024.0
-        return '{:.1f} {}{}'.format(num, 'Yi', suffix)
\ No newline at end of file