import streamlit as st from pages.utils import * from sklearn.svm import SVC from sklearn.pipeline import make_pipeline from sklearn.preprocessing import StandardScaler import joblib from utils import * if 'cleaned_data' not in st.session_state: st.session_state['cleaned_data'] ='' if 'sentences_train' not in st.session_state: st.session_state['sentences_train'] ='' if 'sentences_test' not in st.session_state: st.session_state['sentences_test'] ='' if 'labels_train' not in st.session_state: st.session_state['labels_train'] ='' if 'labels_test' not in st.session_state: st.session_state['labels_test'] ='' if 'svm_classifier' not in st.session_state: st.session_state['svm_classifier'] ='' tab_titles = ['Data Preprocessing', 'Model Training', 'Model Evaluation',"Save Model"] tabs = st.tabs(tab_titles) with tabs[0]: st.header('Data Preprocessing') # Upload the CSV file data = st.file_uploader("Upload CSV file",type="csv") button = st.button("Load data",key="data") if button: with st.spinner('Loading...'): our_data=read_data(data) embeddings=get_embeddings() st.session_state['cleaned_data'] = create_embeddings(our_data,embeddings) st.success('Data Preprocessed') with tabs[1]: st.header('Model Training') button = st.button("Train model",key="model") if button: with st.spinner('Loading...'): st.session_state['sentences_train'], st.session_state['sentences_test'], st.session_state['labels_train'], st.session_state['labels_test']=split_train_test__data(st.session_state['cleaned_data']) st.session_state['svm_classifier'] = make_pipeline(StandardScaler(), SVC(class_weight='balanced')) # fit the support vector machine st.session_state['svm_classifier'].fit(st.session_state['sentences_train'], st.session_state['labels_train']) st.success('Model Trained') with tabs[2]: st.header('Model Evaluation') button = st.button("Evaluate model",key="Evaluation") if button: with st.spinner('Loading...'): accuracy_score=get_score(st.session_state['svm_classifier'],st.session_state['sentences_test'],st.session_state['labels_test']) st.success(f"Validation accuracy is {100*accuracy_score}%!") with tabs[3]: st.header('Save classified model') button = st.button("Save model",key="save") if button: with st.spinner('Loading...'): joblib.dump(st.session_state['svm_classifier'], 'modelsvm.pk1') st.success('Done!')