File size: 7,567 Bytes
267744b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
<<<<<<< HEAD
import pandas as pd
import torch
from datasets import load_dataset, Dataset
from transformers import AutoTokenizer, AutoModelForSequenceClassification, TrainingArguments, Trainer
import numpy as np
from sklearn.metrics import accuracy_score, f1_score, roc_auc_score, classification_report

# Load dataset
dataset = load_dataset("go_emotions")

# Print dataset columns
print("Dataset Columns Before Preprocessing:", dataset["train"].column_names)

# Ensure labels exist
if "labels" not in dataset["train"].column_names:
    raise KeyError("Column 'labels' is missing! Check dataset structure.")

# Load tokenizer
model_checkpoint = "distilbert-base-uncased"

tokenizer = AutoTokenizer.from_pretrained(model_checkpoint)

# Preprocessing function (Take only the first label for single-label classification)
def preprocess_data(batch):
    encoding = tokenizer(batch["text"], padding="max_length", truncation=True)
    
    # Take only the first label (for single-label classification)
    encoding["labels"] = batch["labels"][0] if batch["labels"] else 0  # Default to 0 if empty
    return encoding

# Tokenize dataset
encoded_dataset = dataset.map(preprocess_data, batched=False, remove_columns=["text"])

# Set format for PyTorch
encoded_dataset.set_format("torch")

# Load model for single-label classification (28 classes)
num_labels = 28  # Change based on dataset labels
model = AutoModelForSequenceClassification.from_pretrained(model_checkpoint, num_labels=num_labels)

# Training arguments
args = TrainingArguments(
    output_dir="./results",
    eval_strategy="epoch",
    save_strategy="epoch",
    save_total_limit=1,
    logging_strategy="no",
    per_device_train_batch_size=32,  # Increase batch size
    per_device_eval_batch_size=32,  
    num_train_epochs=2,  # Reduce epochs
    weight_decay=0.01,
    load_best_model_at_end=True,
    fp16=True,  # Mixed precision for speedup
    gradient_accumulation_steps=2,  # Helps with large batch sizes
)


# Compute metrics function
def compute_metrics(eval_pred):
    logits, labels = eval_pred
    
    # Convert logits to class predictions
    predictions = np.argmax(logits, axis=-1)

    accuracy = accuracy_score(labels, predictions)
    f1 = f1_score(labels, predictions, average="weighted")
    
    return {"accuracy": accuracy, "f1": f1}

# Initialize Trainer
trainer = Trainer(
    model=model,
    args=args,
    train_dataset=encoded_dataset["train"],
    eval_dataset=encoded_dataset["validation"],
    compute_metrics=compute_metrics
)

# Train model
trainer.train()
print("Training completed!")

# Save model and tokenizer
model.save_pretrained("./saved_model")
tokenizer.save_pretrained("./saved_model")
print("Model and tokenizer saved!")

# ====== Evaluation on Test Set ======
print("\nEvaluating model on test set...")

# Get test dataset
test_dataset = encoded_dataset["test"]

# Make predictions
predictions = trainer.predict(test_dataset)
logits = predictions.predictions

# Convert logits to class predictions
y_pred = np.argmax(logits, axis=-1)
y_true = test_dataset["labels"].numpy()

# Compute accuracy and F1-score
accuracy = accuracy_score(y_true, y_pred)
f1 = f1_score(y_true, y_pred, average="weighted")

# Print evaluation results
print("\nEvaluation Results:")
print(f"Test Accuracy: {accuracy:.4f}")
print(f"Test F1 Score: {f1:.4f}")

# Print classification report
print("\nClassification Report:\n", classification_report(y_true, y_pred))

# Save test results
pd.DataFrame({"true_labels": y_true.tolist(), "predicted_labels": y_pred.tolist()}).to_csv("test_results.csv", index=False)
print("Test results saved to 'test_results.csv'!")
=======
import pandas as pd
import torch
from datasets import load_dataset, Dataset
from transformers import AutoTokenizer, AutoModelForSequenceClassification, TrainingArguments, Trainer
import numpy as np
from sklearn.metrics import accuracy_score, f1_score, roc_auc_score, classification_report

# Load dataset
dataset = load_dataset("go_emotions")

# Print dataset columns
print("Dataset Columns Before Preprocessing:", dataset["train"].column_names)

# Ensure labels exist
if "labels" not in dataset["train"].column_names:
    raise KeyError("Column 'labels' is missing! Check dataset structure.")

# Load tokenizer
model_checkpoint = "distilbert-base-uncased"

tokenizer = AutoTokenizer.from_pretrained(model_checkpoint)

# Preprocessing function (Take only the first label for single-label classification)
def preprocess_data(batch):
    encoding = tokenizer(batch["text"], padding="max_length", truncation=True)
    
    # Take only the first label (for single-label classification)
    encoding["labels"] = batch["labels"][0] if batch["labels"] else 0  # Default to 0 if empty
    return encoding

# Tokenize dataset
encoded_dataset = dataset.map(preprocess_data, batched=False, remove_columns=["text"])

# Set format for PyTorch
encoded_dataset.set_format("torch")

# Load model for single-label classification (28 classes)
num_labels = 28  # Change based on dataset labels
model = AutoModelForSequenceClassification.from_pretrained(model_checkpoint, num_labels=num_labels)

# Training arguments
args = TrainingArguments(
    output_dir="./results",
    eval_strategy="epoch",
    save_strategy="epoch",
    save_total_limit=1,
    logging_strategy="no",
    per_device_train_batch_size=32,  # Increase batch size
    per_device_eval_batch_size=32,  
    num_train_epochs=2,  # Reduce epochs
    weight_decay=0.01,
    load_best_model_at_end=True,
    fp16=True,  # Mixed precision for speedup
    gradient_accumulation_steps=2,  # Helps with large batch sizes
)


# Compute metrics function
def compute_metrics(eval_pred):
    logits, labels = eval_pred
    
    # Convert logits to class predictions
    predictions = np.argmax(logits, axis=-1)

    accuracy = accuracy_score(labels, predictions)
    f1 = f1_score(labels, predictions, average="weighted")
    
    return {"accuracy": accuracy, "f1": f1}

# Initialize Trainer
trainer = Trainer(
    model=model,
    args=args,
    train_dataset=encoded_dataset["train"],
    eval_dataset=encoded_dataset["validation"],
    compute_metrics=compute_metrics
)

# Train model
trainer.train()
print("Training completed!")

# Save model and tokenizer
model.save_pretrained("./saved_model")
tokenizer.save_pretrained("./saved_model")
print("Model and tokenizer saved!")

# ====== Evaluation on Test Set ======
print("\nEvaluating model on test set...")

# Get test dataset
test_dataset = encoded_dataset["test"]

# Make predictions
predictions = trainer.predict(test_dataset)
logits = predictions.predictions

# Convert logits to class predictions
y_pred = np.argmax(logits, axis=-1)
y_true = test_dataset["labels"].numpy()

# Compute accuracy and F1-score
accuracy = accuracy_score(y_true, y_pred)
f1 = f1_score(y_true, y_pred, average="weighted")

# Print evaluation results
print("\nEvaluation Results:")
print(f"Test Accuracy: {accuracy:.4f}")
print(f"Test F1 Score: {f1:.4f}")

# Print classification report
print("\nClassification Report:\n", classification_report(y_true, y_pred))

# Save test results
pd.DataFrame({"true_labels": y_true.tolist(), "predicted_labels": y_pred.tolist()}).to_csv("test_results.csv", index=False)
print("Test results saved to 'test_results.csv'!")
>>>>>>> b1313c5d084e410cadf261f2fafd8929cb149a4f