app.py

from transformers import AutoTokenizer, AutoModelForCausalLM, TextDataset, DataCollatorForLanguageModeling, Trainer, TrainingArguments
import torch
import streamlit as st

# Load the pre-trained tokenizer and model
tokenizer = AutoTokenizer.from_pretrained("togethercomputer/GPT-NeoXT-Chat-Base-20B")
model = AutoModelForCausalLM.from_pretrained("togethercomputer/GPT-NeoXT-Chat-Base-20B")


dataa = "My name is youssef khemiri i am 21 years old and i am a data scientist"

# Prepare the dataset
train_dataset = TextDataset(
    tokenizer=tokenizer,
    file_path=dataa,
    block_size=128,
)

# Prepare the data collator
data_collator = DataCollatorForLanguageModeling(
    tokenizer=tokenizer, mlm=False,
)

# Initialize the trainer
training_args = TrainingArguments(
    # output_dir='./results',          # output directory
    num_train_epochs=3,              # total number of training epochs
    per_device_train_batch_size=16,  # batch size per device during training
    save_steps=10_000,              # number of steps between saving checkpoints
    save_total_limit=2,             # limit the total amount of checkpoints to save
    prediction_loss_only=True,
    learning_rate=5e-5,
)

trainer = Trainer(
    model=model,
    args=training_args,
    train_dataset=train_dataset,
    data_collator=data_collator,
)

# Fine-tune the model
trainer.train()
st.write("finished training")

# infer
inputs = tokenizer("<human>: Tell me about youssef khemiri\n<bot>:", return_tensors='pt').to(model.device)
outputs = model.generate(**inputs, max_new_tokens=10, do_sample=True, temperature=0.8)
output_str = tokenizer.decode(outputs[0])
st.write(output_str)