train_model_hf.py

import pandas as pd
import numpy as np
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score, precision_recall_fscore_support, confusion_matrix
from transformers import DistilBertTokenizer, DistilBertForSequenceClassification, Trainer, TrainingArguments
import torch

# 1. Load and Preprocess Data
df = pd.read_csv('mail_data.csv', names=['Category', 'Message'], header=None, skiprows=1)
df['label'] = df['Category'].map({'ham': 0, 'spam': 1})

train_texts, test_texts, train_labels, test_labels = train_test_split(
    df['Message'].values.tolist(), df['label'].values.tolist(), test_size=0.2, random_state=42, stratify=df['label'].values
)

# 2. Tokenization
tokenizer = DistilBertTokenizer.from_pretrained('distilbert-base-uncased')

train_encodings = tokenizer(train_texts, truncation=True, padding=True, max_length=128)
test_encodings = tokenizer(test_texts, truncation=True, padding=True, max_length=128)

class EmailDataset(torch.utils.data.Dataset):
    def __init__(self, encodings, labels):
        self.encodings = encodings
        self.labels = labels

    def __getitem__(self, idx):
        item = {key: torch.tensor(val[idx]) for key, val in self.encodings.items()}
        item['labels'] = torch.tensor(self.labels[idx])
        return item

    def __len__(self):
        return len(self.labels)

train_dataset = EmailDataset(train_encodings, train_labels)
test_dataset = EmailDataset(test_encodings, test_labels)

# 3. Model and Metrics
model = DistilBertForSequenceClassification.from_pretrained('distilbert-base-uncased', num_labels=2)

def compute_metrics(pred):
    labels = pred.label_ids
    preds = pred.predictions.argmax(-1)
    precision, recall, f1, _ = precision_recall_fscore_support(labels, preds, average='binary')
    acc = accuracy_score(labels, preds)
    return {
        'accuracy': acc,
        'f1': f1,
        'precision': precision,
        'recall': recall
    }

# 4. Training Arguments
training_args = TrainingArguments(
    output_dir='./results',
    num_train_epochs=3,
    per_device_train_batch_size=16,
    per_device_eval_batch_size=64,
    warmup_steps=500,
    weight_decay=0.01,
    logging_dir='./logs',
    logging_steps=10,
    eval_strategy="epoch",
    save_strategy="epoch",
    load_best_model_at_end=True,
)

# 5. Trainer
trainer = Trainer(
    model=model,
    args=training_args,
    train_dataset=train_dataset,
    eval_dataset=test_dataset,
    compute_metrics=compute_metrics,
)

print("Starting training with HF Trainer...")
trainer.train()

# 6. Evaluation
print("Evaluating...")
eval_results = trainer.evaluate()
print(eval_results)

# Final predictions for detailed report
predictions = trainer.predict(test_dataset)
preds = predictions.predictions.argmax(-1)
labels = predictions.label_ids

from sklearn.metrics import classification_report
report = classification_report(labels, preds, target_names=['ham', 'spam'])
cm = confusion_matrix(labels, preds)

with open('results.txt', 'w') as f:
    f.write(f"Final Evaluation Results:\n{eval_results}\n")
    f.write(f"\nClassification Report:\n{report}\n")
    f.write(f"\nConfusion Matrix:\n{cm}\n")

print("Training complete. Results saved to results.txt")