app.py

from datasets import Dataset
from transformers import AutoTokenizer, AutoModelForCausalLM, TrainingArguments, Trainer
from datafile import QL, resume_data_dict
import streamlit as st
from tensorflow import keras
import tensorflow as tf


# Load your custom data


data = []
for i in range(len(QL['labels'])):
    data.append({"question":QL["queries"][i], "answer":resume_data_dict[QL['labels'][i]]})

# Create a Dataset
dataset = Dataset.from_list(data)

# Load the tokenizer
tokenizer = AutoTokenizer.from_pretrained("microsoft/DialoGPT-medium")
if tokenizer.pad_token is None:
    tokenizer.pad_token = tokenizer.eos_token

def preprocess_function(examples):
    inputs = [f"Question: {q}" for q in examples["question"]]
    model_inputs = tokenizer(inputs, padding="max_length", truncation=True, max_length=128)

    # Setup the tokenizer for targets
    with tokenizer.as_target_tokenizer():
        labels = tokenizer(examples["answer"], padding="max_length", truncation=True, max_length=128)

    model_inputs["labels"] = labels["input_ids"]
    return model_inputs

# Apply preprocessing

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


model = AutoModelForCausalLM.from_pretrained("microsoft/DialoGPT-medium")
model.resize_token_embeddings(len(tokenizer))


# Define training arguments
training_args = TrainingArguments(
    output_dir="./resume_bot",
    num_train_epochs=3,
    per_device_train_batch_size=2,
    per_device_eval_batch_size=2,
    warmup_steps=10,
    weight_decay=0.01,
    logging_dir="./logs",
    logging_steps=10,
    save_steps=500,
    evaluation_strategy="steps"
)

# Initialize the Trainer
trainer = Trainer(
    model=model,
    args=training_args,
    train_dataset=tokenized_dataset,
    eval_dataset=tokenized_dataset,
    tokenizer=tokenizer
)

# Train the model

trainer.train()


model_name = "./resume_bot"  # Path to your fine-tuned model
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name)


st.title("Resume Chatbot")

if 'history' not in st.session_state:
    st.session_state.history = []

user_input = st.text_input("You: ", "")

if user_input:
    # Encode the input
    input_ids = tokenizer.encode(user_input + tokenizer.eos_token, return_tensors='pt')

    try:
        response_ids = model.generate(input_ids, max_length=1000, pad_token_id=tokenizer.eos_token_id)
        bot_response = tokenizer.decode(response_ids[0], skip_special_tokens=True)

        # Update the chat history
        st.session_state.history.append(f"You: {user_input}")
        st.session_state.history.append(f"Bot: {bot_response}")

        # Display the bot response
        st.write(f"Bot: {bot_response}")

    except Exception as e:
        st.error(f"Error generating response: {e}")

# Add a button to clear the conversation
if st.button("Reset Conversation"):
    st.session_state.history = []