Blizzer

!pip install transformers datasets torch
!pip install transformers[torch] -U
!pip install accelerate -U

import torch
from transformers import GPT2LMHeadModel, GPT2Tokenizer, DataCollatorForLanguageModeling
from transformers import Trainer, TrainingArguments
from datasets import load_dataset

# Load pre-trained model and tokenizer
model_name = "gpt2"
model = GPT2LMHeadModel.from_pretrained(model_name)
tokenizer = GPT2Tokenizer.from_pretrained(model_name)

# Set padding token
tokenizer.pad_token = tokenizer.eos_token

# Load and preprocess the dataset
dataset = load_dataset("emotion", split="train")

def preprocess_function(examples):
    return tokenizer([f"Emotion: {text}" for text in examples["text"]], truncation=True, padding="max_length", max_length=64)

tokenized_dataset = dataset.map(preprocess_function, batched=True, remove_columns=dataset.column_names)

# Convert to PyTorch tensors
tokenized_dataset.set_format("torch")

# Create DataCollator
data_collator = DataCollatorForLanguageModeling(
    tokenizer=tokenizer,
    mlm=False
)

# Define training arguments
training_args = TrainingArguments(
    output_dir="./gpt2-emotion-finetuned",
    overwrite_output_dir=True,
    num_train_epochs=3,
    per_device_train_batch_size=4,
    save_steps=10_000,
    save_total_limit=2,
)

# Create Trainer
trainer = Trainer(
    model=model,
    args=training_args,
    data_collator=data_collator,
    train_dataset=tokenized_dataset,
)

# Fine-tune the model
trainer.train()

# Save the fine-tuned model
model.save_pretrained("./gpt2-emotion-finetuned")
tokenizer.save_pretrained("./gpt2-emotion-finetuned")

# Load the fine-tuned model and tokenizer
fine_tuned_model = GPT2LMHeadModel.from_pretrained("./gpt2-emotion-finetuned")
fine_tuned_tokenizer = GPT2Tokenizer.from_pretrained("./gpt2-emotion-finetuned")

# បង្កើតអត្ថបទ
prompt = "Emotion: i didnt feel well"
input_ids = fine_tuned_tokenizer.encode(prompt, return_tensors="pt")
output = fine_tuned_model.generate(input_ids, max_length=100, num_return_sequences=1, no_repeat_ngram_size=2)

generated_text = fine_tuned_tokenizer.decode(output[0], skip_special_tokens=True)
print(generated_text)

Output
Emotion: i didnt feel well enough to go to the doctor but i was feeling good enough that i could go for a walk and not feel so bad about it all the time and that was good for me too because i had been feeling pretty good about my health for the past week and a half and i just needed to get through it without feeling like i got a bad grade or something and then i would be fine again and again until i went to my doctor and got my results and it was