File size: 4,948 Bytes
11650ac 288fd1b 11650ac 288fd1b 11650ac 288fd1b 11650ac 288fd1b 11650ac 288fd1b 11650ac 288fd1b 11650ac 288fd1b 11650ac 288fd1b 11650ac 288fd1b 11650ac 288fd1b 11650ac |
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 |
# 1. IMPORT NECESSARY LIBRARIES
import streamlit as st
from transformers import pipeline
# 2. SET UP THE PAGE CONFIGURATION AND TITLE
st.set_page_config(
page_title="GUVI AI Assistant",
page_icon="🤖",
layout="wide",
initial_sidebar_state="expanded"
)
# 3. LOAD ALL THREE AI MODELS
@st.cache_resource
def load_models():
"""Loads and returns all three required AI pipelines."""
translator = pipeline("translation", model="facebook/mbart-large-50-many-to-many-mmt")
generator = pipeline("text-generation", model="aswinprasath31/guvi-gpt2-finetuned")
detector = pipeline("text-classification", model="papluca/xlm-roberta-base-language-detection")
return translator, generator, detector
# Load the models. A spinner is shown in the sidebar while loading.
with st.spinner("Initializing AI models, please wait..."):
translator, generator, detector = load_models()
# This dictionary maps short language codes to the longer codes needed by the translator.
MODEL_LANG_MAP = {
"ar": "ar_AR", "cs": "cs_CZ", "de": "de_DE", "en": "en_XX", "es": "es_XX", "et": "et_EE", "fi": "fi_FI",
"fr": "fr_XX", "gu": "gu_IN", "hi": "hi_IN", "it": "it_IT", "ja": "ja_XX", "kk": "kk_KZ", "ko": "ko_KR",
"lt": "lt_LT", "lv": "lv_LV", "my": "my_MM", "ne": "ne_NP", "nl": "nl_XX", "ro": "ro_RO", "ru": "ru_RU",
"si": "si_LK", "tr": "tr_TR", "vi": "vi_VN", "zh": "zh_CN", "af": "af_ZA", "az": "az_AZ", "bn": "bn_IN",
"fa": "fa_IR", "he": "he_IL", "hr": "hr_HR", "id": "id_ID", "ka": "ka_GE", "km": "km_KH", "mk": "mk_MK",
"ml": "ml_IN", "mn": "mn_MN", "mr": "mr_IN", "pl": "pl_PL", "ps": "ps_AF", "pt": "pt_XX", "sv": "sv_SE",
"sw": "sw_KE", "ta": "ta_IN", "te": "te_IN", "th": "th_TH", "tl": "tl_XX", "uk": "uk_UA", "ur": "ur_PK",
"xh": "xh_ZA", "gl": "gl_ES", "sl": "sl_SI"
}
# 4. SIDEBAR CONTENT
with st.sidebar:
st.title("About the Chatbot")
st.info(
"This is an advanced multilingual AI assistant for GUVI. "
"It leverages a fine-tuned GPT-2 model to provide accurate, "
"domain-specific answers about GUVI's courses."
)
st.success(
"**Key Features:**\n"
"- **Auto-Detects Language:** Chat in over 20 languages.\n"
"- **Domain-Specific:** Fine-tuned on GUVI's data.\n"
"- **Real-time:** Get instant answers to your questions."
)
st.markdown("---")
st.header("Tech Stack")
st.markdown(
"""
- Streamlit
- Hugging Face Transformers
- PyTorch
- Google Colab
"""
)
# 5. SET UP THE MAIN CHAT INTERFACE
st.title("GUVI AI Learning Assistant 🤖")
# Initialize chat history in session state
if "messages" not in st.session_state:
st.session_state.messages = [{"role": "assistant", "content": "Hello! How can I help you with GUVI's courses today?"}]
# Display prior chat messages
for message in st.session_state.messages:
with st.chat_message(message["role"]):
st.markdown(message["content"])
# 6. DEFINE THE CHATBOT'S LOGIC
def get_bot_response(user_input):
"""Processes user input through the AI pipeline and returns the final response."""
if len(user_input.split()) < 3:
return "That's a broad topic! Could you please ask a more specific question about it?"
detection_result = detector(user_input)
detected_lang_code = detection_result[0]['label']
src_lang_code = MODEL_LANG_MAP.get(detected_lang_code, "en_XX")
if src_lang_code == "en_XX":
english_query = user_input
else:
translated_to_en = translator(user_input, src_lang=src_lang_code, tgt_lang="en_XX")
english_query = translated_to_en[0]['translation_text']
prompt = f"Answer the following question about GUVI's online courses in a helpful and clear manner: {english_query}"
gpt_response = generator(prompt, max_length=200, num_return_sequences=1)
english_response = gpt_response[0]['generated_text']
if src_lang_code == "en_XX":
final_response = english_response
else:
translated_to_orig = translator(english_response, src_lang="en_XX", tgt_lang=src_lang_code)
final_response = translated_to_orig[0]['translation_text']
return final_response
# 7. ACCEPT AND PROCESS USER INPUT
if prompt := st.chat_input("Ask me about Data Science, Python, or other courses..."):
st.session_state.messages.append({"role": "user", "content": prompt})
with st.chat_message("user"):
st.markdown(prompt)
with st.chat_message("assistant"):
with st.spinner("Thinking..."):
response = get_bot_response(prompt)
st.markdown(response)
st.session_state.messages.append({"role": "assistant", "content": response})
# 8. ADD THE FOOTER
st.markdown("---")
st.markdown(
"""
<div style="text-align: center; color: grey;">
Developed by <b>Aswinprasath V</b> | Supported by <b>GUVI</b>
</div>
""",
unsafe_allow_html=True
) |