Upload 3 files

pages/answers.py

@@ -0,0 +1,35 @@
+import streamlit as st
+from transformers import AutoModelForQuestionAnswering, AutoTokenizer, pipeline
+
+model_name = "timpal0l/mdeberta-v3-base-squad2"
+model = AutoModelForQuestionAnswering.from_pretrained(model_name)
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+
+def get_answer(context, question):
+    nlp = pipeline('question-answering', model=model, tokenizer=tokenizer)
+    QA_input = {'question': question, 'context': context}
+    res = nlp(QA_input)
+    answer = res['answer']
+    return answer
+
+def main():
+    st.title("Question Answering App")
+    st.markdown("Enter the context and question, then click on 'Get Answer' to retrieve the answer.")
+
+    
+    context = st.text_area("Context", "Enter the context here...")
+    question = st.text_input("Question", "Enter the question here...")
+
+    
+    if st.button("Get Answer"):
+        
+        if context.strip() == "" or question.strip() == "":
+            st.warning("Please enter the context and question.")
+        else:
+            
+            answer = get_answer(context, question)
+            st.success(f"Answer: {answer}")
+
+
+if __name__ == "__main__":
+    main()

pages/gpt.py

@@ -0,0 +1,35 @@
+from transformers import GPT2LMHeadModel, GPT2Tokenizer
+import torch
+
+model = GPT2LMHeadModel.from_pretrained(
+    'sberbank-ai/rugpt3small_based_on_gpt2',
+    output_attentions = False,
+    output_hidden_states = False,
+)
+# Вешаем сохраненные веса на нашу модель
+model.load_state_dict(torch.load('models/model.pt'), map_location=torch.device('cpu'))
+
+
+def generate_text(model, tokenizer, prompt, length, num_samples, temperature):
+    input_ids = tokenizer.encode(prompt, return_tensors='pt')
+    output_sequences = model.generate(
+        input_ids=input_ids,
+        max_length=length,
+        num_return_sequences=num_samples,
+        temperature=temperature
+    )
+
+    generated_texts = []
+    for output_sequence in output_sequences:
+        generated_text = tokenizer.decode(output_sequence, clean_up_tokenization_spaces=True)
+        generated_texts.append(generated_text)
+
+    return generated_texts
+
+
+if st.button('Сгенерировать текст'):
+    generated_texts = generate_text(model, tokenizer, prompt, length, num_samples, temperature)
+    for i, text in enumerate(generated_texts):
+        st.write(f'Текст {i+1}:')
+        st.write(text)
+

pages/imdb.py

@@ -0,0 +1,19 @@
+import os
+import numpy as np
+import pandas as pd
+
+import matplotlib.pyplot as plt
+
+import re
+import string
+from collections import Counter
+from nltk.corpus import stopwords
+stop_words = set(stopwords.words('english'))
+
+from sklearn.model_selection import train_test_split
+
+import torch
+from torch.utils.data import DataLoader, TensorDataset
+import torch.nn as nn
+import torchutils as tu
+from torchmetrics.classification import BinaryAccuracy