add text generator

app.py

@@ -1,47 +1,3 @@
-import pickle
-
 import streamlit as st
 
-from preprocessing import data_preprocessing
-
-# Load preprocessing steps
-with open("vectorizer.pkl", "rb") as f:
-    vectorizer = pickle.load(f)
-
-# Load trained model
-with open("logreg_model.pkl", "rb") as f:
-    logreg = pickle.load(f)
-
-
-# Define function for preprocessing input text
-def preprocess_text(text):
-    # Apply preprocessing steps (cleaning, tokenization, vectorization)
-    clean_text = data_preprocessing(
-        text
-    )  # Assuming data_preprocessing is your preprocessing function
-    print("Clean text ", clean_text)
-    vectorized_text = vectorizer.transform([" ".join(clean_text)])
-    return vectorized_text
-
-
-# Define function for making predictions
-def predict_sentiment(text):
-    # Preprocess input text
-    processed_text = preprocess_text(text)
-    print(preprocess_text)
-    # Make prediction
-    prediction = logreg.predict(processed_text)
-    return prediction
-
-
-# Streamlit app code
 st.title("Sentiment Analysis with Logistic Regression")
-text_input = st.text_input("Enter your review:")
-if st.button("Predict"):
-    prediction = predict_sentiment(text_input)
-    if prediction == 1:
-        st.write("prediction")
-        st.write("Отзыв положительный")
-    elif prediction == 0:
-        st.write("prediction")
-        st.write("Отзыв отрицательный")

finetuned_model/config.json

@@ -0,0 +1,41 @@
+{
+  "_name_or_path": "sberbank-ai/rugpt3small_based_on_gpt2",
+  "activation_function": "gelu_new",
+  "architectures": [
+    "GPT2LMHeadModel"
+  ],
+  "attn_pdrop": 0.1,
+  "bos_token_id": 1,
+  "embd_pdrop": 0.1,
+  "eos_token_id": 2,
+  "gradient_checkpointing": false,
+  "id2label": {
+    "0": "LABEL_0"
+  },
+  "initializer_range": 0.02,
+  "label2id": {
+    "LABEL_0": 0
+  },
+  "layer_norm_epsilon": 1e-05,
+  "model_type": "gpt2",
+  "n_ctx": 2048,
+  "n_embd": 768,
+  "n_head": 12,
+  "n_inner": null,
+  "n_layer": 12,
+  "n_positions": 2048,
+  "pad_token_id": 0,
+  "reorder_and_upcast_attn": false,
+  "resid_pdrop": 0.1,
+  "scale_attn_by_inverse_layer_idx": false,
+  "scale_attn_weights": true,
+  "summary_activation": null,
+  "summary_first_dropout": 0.1,
+  "summary_proj_to_labels": true,
+  "summary_type": "cls_index",
+  "summary_use_proj": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.38.2",
+  "use_cache": true,
+  "vocab_size": 50264
+}

finetuned_model/generation_config.json

@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "pad_token_id": 0,
+  "transformers_version": "4.38.2"
+}

finetuned_model/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fcdf8d066aa4a05109a1867faf91ab3645bfcec52881d0a9572992c20fbe3120
+size 500941440

pages/review_predictor.py

@@ -0,0 +1,58 @@
+import pickle
+
+import streamlit as st
+
+from preprocessing import data_preprocessing
+
+# Load preprocessing steps
+with open("vectorizer.pkl", "rb") as f:
+    logreg_vectorizer = pickle.load(f)
+
+# Load trained model
+with open("logreg_model.pkl", "rb") as f:
+    logreg_predictor = pickle.load(f)
+
+
+# Define function for preprocessing input text
+@st.cache
+def preprocess_text(text):
+    # Apply preprocessing steps (cleaning, tokenization, vectorization)
+    clean_text = data_preprocessing(
+        text
+    )  # Assuming data_preprocessing is your preprocessing function
+    print("Clean text ", clean_text)
+    vectorized_text = vectorizer.transform([" ".join(clean_text)])
+    return vectorized_text
+
+
+# Define function for making predictions
+@st.cache
+def predict_sentiment(text):
+    # Preprocess input text
+    processed_text = preprocess_text(text)
+    # Make prediction
+    prediction = logreg_predictor.predict(processed_text)
+    return prediction
+
+
+st.sidebar.title("Model Selection")
+model_type = st.sidebar.radio("Select Model Type", ["Classic ML", "LSTM", "BERT"])
+st.title("Review Prediction")
+
+# Streamlit app code
+st.title("Sentiment Analysis with Logistic Regression")
+text_input = st.text_input("Enter your review:")
+if st.button("Predict"):
+    if model_type == "Classic ML":
+        prediction = predict_sentiment(text_input)
+    elif model_type == "LSTM":
+        prediction = 1
+    elif model_type == "BERT":
+        prediction = 1
+
+    if prediction == 1:
+        st.write("prediction")
+        st.write("Отзыв положительный")
+    elif prediction == 0:
+        st.write("prediction")
+        st.write("Отзыв отрицательный")

pages/text_generator.py

@@ -0,0 +1,27 @@
+import streamlit as st
+import torch
+from transformers import GPT2LMHeadModel, GPT2Tokenizer
+
+model_path = "finetuned_model/"
+model_name = "sberbank-ai/rugpt3small_based_on_gpt2"
+tokenizer = GPT2Tokenizer.from_pretrained(model_name)
+model = GPT2LMHeadModel.from_pretrained(model_path)
+
+
+promt = st.text_input("Ask a question")
+generate = st.button("Generate")
+if generate:
+    if not promt:
+        st.write("42")
+    promt = tokenizer.encode(promt, return_tensors="pt")
+    model.eval()
+    with torch.no_grad():
+        out = model.generate(
+            promt,
+            do_sample=True,
+            num_beams=2,
+            temperature=1.5,
+            top_p=0.9,
+        )
+    out = list(map(tokenizer.decode, out))[0]
+    st.write(out)