Spaces:
Sleeping
Sleeping
| # import streamlit as st | |
| # from transformers import AutoModelForCausalLM, AutoTokenizer | |
| # # Load the model and tokenizer | |
| # @st.cache_resource | |
| # def load_model_and_tokenizer(): | |
| # model_name = "microsoft/DialoGPT-medium" # Replace with your chosen model | |
| # tokenizer = AutoTokenizer.from_pretrained(model_name) | |
| # model = AutoModelForCausalLM.from_pretrained(model_name) | |
| # return tokenizer, model | |
| # tokenizer, model = load_model_and_tokenizer() | |
| # # Streamlit App | |
| # st.title("General Chatbot") | |
| # st.write("A chatbot powered by an open-source model from Hugging Face.") | |
| # # Initialize the conversation | |
| # if "conversation_history" not in st.session_state: | |
| # st.session_state["conversation_history"] = [] | |
| # # Input box for user query | |
| # user_input = st.text_input("You:", placeholder="Ask me anything...", key="user_input") | |
| # if st.button("Send") and user_input: | |
| # # Append user input to history | |
| # st.session_state["conversation_history"].append({"role": "user", "content": user_input}) | |
| # # Tokenize and generate response | |
| # input_ids = tokenizer.encode(user_input + tokenizer.eos_token, return_tensors="pt") | |
| # chat_history_ids = model.generate(input_ids, max_length=1000, pad_token_id=tokenizer.eos_token_id) | |
| # response = tokenizer.decode(chat_history_ids[:, input_ids.shape[-1]:][0], skip_special_tokens=True) | |
| # # Append model response to history | |
| # st.session_state["conversation_history"].append({"role": "assistant", "content": response}) | |
| # # Display the conversation | |
| # for message in st.session_state["conversation_history"]: | |
| # if message["role"] == "user": | |
| # st.write(f"**You:** {message['content']}") | |
| # else: | |
| # st.write(f"**Bot:** {message['content']}") | |
| import streamlit as st | |
| from transformers import pipeline, AutoTokenizer, AutoModelForSeq2SeqLM | |
| st.title("🤖 Smart Chatbot") | |
| def load_model(): | |
| model_name = "facebook/blenderbot-3B" | |
| tokenizer = AutoTokenizer.from_pretrained(model_name) | |
| model = AutoModelForSeq2SeqLM.from_pretrained(model_name) | |
| return pipeline("text2text-generation", model=model, tokenizer=tokenizer) | |
| chatbot = load_model() | |
| if "conversation" not in st.session_state: | |
| st.session_state.conversation = [] | |
| # Display chat history | |
| for msg in st.session_state.conversation: | |
| with st.chat_message(msg["role"]): | |
| st.markdown(msg["content"]) | |
| if prompt := st.chat_input("Ask me anything"): | |
| # Add user message | |
| st.session_state.conversation.append({"role": "user", "content": prompt}) | |
| # Format context | |
| context = "\n".join([f"{msg['role']}: {msg['content']}" for msg in st.session_state.conversation[-3:]]) | |
| try: | |
| with st.spinner("Thinking..."): | |
| response = chatbot( | |
| context, | |
| max_length=200, | |
| temperature=0.9, | |
| top_k=60, | |
| top_p=0.9, | |
| num_beams=5, | |
| no_repeat_ngram_size=3 | |
| )[0]['generated_text'] | |
| # Clean response | |
| response = response.split("assistant:")[-1].strip() | |
| # Ensure meaningful response | |
| if not response or response.lower() in ["i don't know", "i'm not sure"]: | |
| response = "I need to learn more about that. Could you clarify?" | |
| except Exception as e: | |
| response = "Let me check my knowledge sources and get back to you on that." | |
| st.session_state.conversation.append({"role": "assistant", "content": response}) | |
| st.rerun() |