app.py

import torch
import pandas as pd
from transformers import pipeline, AutoTokenizer, AutoModelForCausalLM, AutoModelForSequenceClassification, TrainingArguments, Trainer
import gradio as gr
from gradio.mix import Parallel, Series
#import torch.nn.functional as F
from aitextgen import aitextgen

#IMPLEMENT FINE TUNING

#device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')

#from datasets import load_dataset
#dataset = load_dataset("bananabot/engMollywoodSummaries")
#tokenizer = AutoTokenizer.from_pretrained("EleutherAI/gpt-neo-1.3B")

#tokenizer.pad_token = tokenizer.eos_token

#def tokenize_function(examples):
#    return tokenizer(examples["text"], padding="max_length", truncation=True)
#tokenized_datasets = dataset.map(tokenize_function, batched=True)

#model = AutoModelForCausalLM.from_pretrained("EleutherAI/gpt-neo-1.3B").to(device)
#training_args = TrainingArguments(output_dir="test_trainer")
#small_train_dataset = tokenized_datasets["train"].shuffle(seed=42).select(range(1000))
#small_eval_dataset = tokenized_datasets["test"].shuffle(seed=42).select(range(1000))
#def compute_metrics(eval_pred):
#    logits, labels = eval_pred
#    predictions = np.argmax(logits, axis=-1)
#    return metric.compute(predictions=predictions, references=labels)

#trainer = Trainer(
#    model=model,
#    args=training_args,
#    train_dataset=small_train_dataset,
#    eval_dataset=small_eval_dataset,
#    compute_metrics=compute_metrics,
#)

#trainer.train()




ai = aitextgen(model="EleutherAI/gpt-neo-125M")

#model_name = "EleutherAI/gpt-neo-125M"
#tokenizer = AutoTokenizer.from_pretrained(model_name)


#max_length=123
#input_txt = "This malayalam movie is about"
#n_steps = 8

#input_ids = tokenizer(input_txt, return_tensors="pt")["input_ids"].to(device)
#output = model.generate(input_ids,  max_length=max_length, num_beams=5, do_sample=True, no_repeat_ngram_size=2, temperature=1.37, top_k=69, top_p=0.96)
#print(tokenizer.decode(output[0]))

#def generate(input_txt):
#    output = model.generate(input_ids,  max_length=max_length, num_beams=5, do_sample=True, no_repeat_ngram_size=2, temperature=1.37,           top_k=69, top_p=0.96)
#    print (output)

def ai_text(inp):
  generated_text = ai.generate_one(max_length=234, prompt = inp, no_repeat_ngram_size=3, num_beams=7, do_sample=True, temperature=1.37, top_k=69, top_p=0.96)
  print(type(generated_text))
  return generated_text
    
#inputs= gr.inputs.Textbox(lines=7, placeholder="Enter the beginning of your mollywood movie idea and the നിർമ്മിത ബുദ്ധി will fill in the rest...")

#IMPLEMENT KURIAN BENOY

#generator = output 
#translator = gr.Interface.load("models/Helsinki-NLP/opus-mt-en-ml")
#gr.Series(generator, translator, inputs=gr.inputs.Textbox(lines=7, label="Input Text")).launch()  # this demo generates text, then translates it to Malayalam, and outputs the final result.

#interface = gr.Interface(fn=generate, 
#                        inputs=inputs,
#                        outputs='text',
#                        title='AI based Mollywood movie idea generator')

#interface.launch()

output_text = gr.outputs.Textbox()
gr.Interface(ai_text,"textbox", output_text, title="AI based Mollywood movie idea generator",
             description="Enter the beginning of your malayalam movie idea and the നിർമ്മിത ബുദ്ധി will fill in the rest...").launch()