Update app.py

app.py

@@ -3,16 +3,22 @@ import torch
 from datasets import load_dataset
 from console_logging.console import Console
 import numpy as np
+from transformers import AutoModelForSequenceClassification, AutoTokenizer
+from transformers import TrainingArguments, Trainer
+from sklearn.metrics import f1_score, roc_auc_score, accuracy_score
+from transformers import EvalPrediction
+import torch
 console = Console()
 
 dataset = load_dataset("zeroshot/twitter-financial-news-sentiment", )
 
-from transformers import AutoModelForSequenceClassification, AutoTokenizer
 
 model = AutoModelForSequenceClassification.from_pretrained("bert-base-uncased")
 tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased")
 
-labels = [label for label in dataset['train'].features.keys() if label not in ['text']]
+#labels = [label for label in dataset['train'].features.keys() if label not in ['text']]
+
+labels = ["Bearish", "Bullish", "Neutral"]
 
 def preprocess_data(examples):
   # take a batch of texts
@@ -20,7 +26,22 @@ def preprocess_data(examples):
   # encode them
   encoding = tokenizer(text, padding="max_length", truncation=True, max_length=128)
   # add labels
-  labels_batch = {k: examples[k] for k in examples.keys() if k in labels}
+  #labels_batch = {k: examples[k] for k in examples.keys() if k in labels}
+  labels_batch = {'Bearish': [], 'Bullish': [], 'Neutral': []}
+  for i in range (len(examples['label'])):
+    labels_batch["Bearish"].append(False)
+    labels_batch["Bullish"].append(False)
+    labels_batch["Neutral"].append(False)
+    
+    if examples['label'][i] == 0:
+      labels_batch["Bearish"][i] = True
+
+    elif examples['label'][i] == 1:
+      labels_batch["Bullish"][i] = True
+
+    else:
+      labels_batch["Neutral"][i] = True
+
   # create numpy array of shape (batch_size, num_labels)
   labels_matrix = np.zeros((len(text), len(labels)))
   # fill numpy array
@@ -38,7 +59,8 @@ encoded_dataset.set_format("torch")
 id2label = {idx:label for idx, label in enumerate(labels)}
 label2id = {label:idx for idx, label in enumerate(labels)}
 
-model = AutoModelForSequenceClassification.from_pretrained("bert-base-uncased", 
+model = AutoModelForSequenceClassification.from_pretrained("bert-base-uncased",
+                                                           problem_type="multi_label_classification", 
                                                            num_labels=len(labels),
                                                            id2label=id2label,
                                                            label2id=label2id)
@@ -46,8 +68,6 @@ model = AutoModelForSequenceClassification.from_pretrained("bert-base-uncased",
 batch_size = 8
 metric_name = "f1"
 
-from transformers import TrainingArguments, Trainer
-
 args = TrainingArguments(
     f"bert-finetuned-sem_eval-english",
     evaluation_strategy = "epoch",
@@ -62,11 +82,6 @@ args = TrainingArguments(
     #push_to_hub=True,
 )
 
-
-from sklearn.metrics import f1_score, roc_auc_score, accuracy_score
-from transformers import EvalPrediction
-import torch
-    
 # source: https://jesusleal.io/2021/04/21/Longformer-multilabel-classification/
 def multi_label_metrics(predictions, labels, threshold=0.5):
     # first, apply sigmoid on predictions which are of shape (batch_size, num_labels)
@@ -106,22 +121,4 @@ trainer = Trainer(
 
 trainer.train()
 
-trainer.evaluate()
-
-"""
-
-categories = ('Car in good condition','Damaged Car')
-
-def is_car(x) : return x[0].isupper()
-
-def image_classifier(img):
-    pred,index,probs = learn.predict(img) 
-    return dict(zip(categories, map(float,probs)))
-
-# image = gr.inputs.Image(shape=(192,192))
-image = gr.components.Image(shape=(192,192))
-label = gr.components.Label()
-examples = ['./car.jpg','./crash.jpg','./carf.jpg']
-
-intf = gr.Interface(fn= image_classifier,inputs=image,outputs=label,examples=examples)
-intf.launch()"""
+trainer.evaluate()