File size: 4,610 Bytes
5ecde30 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 |
import streamlit as st
'---'
#st.title('Watson Assistant VDF TOBi improvement')
st.markdown(
"""
<style>
.stTextInput > div > div > input {
background-color: #d3d3d3;
}
body {
background-color: #f0f0f0;
}
</style>
""",
unsafe_allow_html=True
)
st.header('Watson Assistant VDF TOBi improvement')
'---'
st.write('The model is trained on the TOBi 🤖 intents in Romanian language.')
#st.write('🤖')
#:robot_face:
import os
import pandas as pd
import re
from time import time
from src.E_Model_utils import load_model, train_model, get_embeddings
from src.E_Faiss_utils import load_embeddings_and_index, normalize_embeddings
from src.A_Preprocess import load_data, clean_text
import warnings
warnings.filterwarnings("ignore", category=FutureWarning)
model_name = st.sidebar.radio("Selectează modelul 👇", ["other","e5_small_fine_tuned_model","multilingual-e5-small","all-MiniLM-L6-v2","all-distilroberta-v1"])
# Load the saved embeddings
#model_name = "xlm-roberta-base" # Choose the desired model
#model_name = "xlm-r-distilroberta-base-paraphrase-v1"
# Model path
# Load the trained model
if model_name != "other":
# future improvement: add a loading spinner
model_path = f"output/fine-tuned-model"
st.write("Model path:", model_path)
#model = load_model(model_path)
if model_name == "multilingual-e5-small":
infloat_model_name = "intfloat/multilingual-e5-small"
model = load_model(infloat_model_name)
elif model_name == "e5_small_fine_tuned_model":
infloat_model_name = "intfloat/multilingual-e5-small"
model = load_model(infloat_model_name)
pass#model = load_model(model_path)
else:
model = load_model(model_name)
st.write(f"Modelul selectat: {model_name}")
st.write("Model loaded successfully!")
# Load the embeddings and the index
#embeddings, index = load_embeddings_and_index(f"embeddings/{model_name}_vector_db.index")
st.stop()
# Upload the intents data
uploaded_file = st.file_uploader("Încarcă fișierul cu intenții", type="csv")
if uploaded_file is not None:
data = pd.read_csv(uploaded_file)
st.write("CSV file successfully uploaded!")
# Save data to session state
st.session_state.data = data
else:
# If no file is uploaded, try to load data from session state
data = st.session_state.data if 'data' in st.session_state else None
if data is not None:
# Extract utterances and intents
utterances = data['utterance'].tolist()
intents = data['intent'].tolist()
user_text = st.text_input("Te rog introdu un text.")
if user_text:
if st.button("Identifică Intenția"):
start = time()
st.write("Procesare text...")
st.write(start)
cleaned_text = clean_text(user_text)
input_embedding = get_embeddings(model, [cleaned_text])
normalized_embedding = normalize_embeddings(input_embedding)
embeddings, index = load_embeddings_and_index(f"embeddings/{model_name}_vector_db.index")
D, I = index.search(normalized_embedding, 1) # Caută cel mai apropiat vecin
print(I)
intents = data['intent'].tolist()
intent = intents[I[0][0]]
distance = D[0][0]
similarity = 1 / (1 + distance)
st.write(f"Intenția identificată: {intent}")
st.write(f"Nivel de încredere: {similarity:.4f}")
st.write(f"Timp de răspuns: {time() - start:.4f} secunde")
else:
st.write("Te rog introdu un text.")
st.stop()
# Endpoint pentru identificarea intenției
input_text = st.text_input("Introdu mai jos textul! 👇", label_visibility="visible")
if input_text:
start = time()
input_embeddings = model.encode([input_text])
if st.button("Identifică Intenția"):
if input_text:
cleaned_text = clean_text(input_text)
input_embedding = get_embeddings(model, [cleaned_text])
normalized_embedding = normalize_embeddings(input_embedding)
D, I = index.search(normalized_embedding, 1) # Caută cel mai apropiat vecin
#print(I)
intents = data['intent'].tolist()
intent = intentions[I[0][0]]
distance = D[0][0]
similarity = 1 / (1 + distance)
st.write(f"Intenția identificată: {intent}")
st.write(f"Nivel de încredere: {similarity:.4f}")
st.write(f"Timp de răspuns: {time() - start:.4f} secunde")
else:
st.write("Te rog introdu un text.")
st.stop()
|