File size: 2,190 Bytes
402354e 9f718b9 402354e 9f718b9 402354e |
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 |
from datasets import load_dataset
from transformers import GPT2LMHeadModel, GPT2Tokenizer
from transformers import Trainer, TrainingArguments
ds = load_dataset("openai/summarize_from_feedback", "axis")
# Load dataset (replace with your actual dataset path or URL)
dataset = load_dataset("path_to_your_dataset")
from datasets import load_dataset
# Load the pre-trained model and tokenizer
model_name = "gpt2" # Or another model like T5, BART, etc.
model = GPT2LMHeadModel.from_pretrained(model_name)
tokenizer = GPT2Tokenizer.from_pretrained(model_name)
# Preprocess the dataset: tokenizing the answers and feedback
def preprocess_function(examples):
return tokenizer(examples['answer'], truncation=True, padding='max_length', max_length=512)
# Apply preprocessing to the dataset
encoded_dataset = dataset.map(preprocess_function, batched=True)
# Define the training arguments
training_args = TrainingArguments(
output_dir="./results", # Where to save the model
evaluation_strategy="epoch", # Evaluate after each epoch
per_device_train_batch_size=4, # Batch size for training
per_device_eval_batch_size=4, # Batch size for evaluation
num_train_epochs=3, # Number of training epochs
logging_dir='./logs', # Directory for logging
logging_steps=10,
save_steps=500, # Save model checkpoint every 500 steps
warmup_steps=200, # Gradual learning rate warm-up steps
weight_decay=0.01, # Weight decay for regularization
load_best_model_at_end=True, # Load the best model based on evaluation metric
)
# Initialize the Trainer
trainer = Trainer(
model=model, # The pre-trained model you want to fine-tune
args=training_args, # The training configuration
train_dataset=encoded_dataset['train'], # Your training data
eval_dataset=encoded_dataset['test'], # Your evaluation data
)
# Start training
trainer.train()
# Save the model after training
model.save_pretrained("./fine_tuned_model")
tokenizer.save_pretrained("./fine_tuned_model")
print("Model fine-tuning complete!")
|