Spaces:
Sleeping
Sleeping
| ''' | |
| Taken directly from : https://huggingface.co/spaces/Sagar23p/mistralAI_chatBoat/tree/main | |
| ''' | |
| import streamlit as st | |
| from huggingface_hub import InferenceClient | |
| import os | |
| import sys | |
| st.title("ChatGPT-like Chatbot") | |
| base_url="https://api-inference.huggingface.co/models/" | |
| API_KEY = os.environ.get('HUGGINGFACE_API_KEY') | |
| # print(API_KEY) | |
| # headers = {"Authorization":"Bearer "+API_KEY} | |
| model_links ={ | |
| "Mistral-7B":base_url+"mistralai/Mistral-7B-Instruct-v0.2", | |
| "Phi-3.5":base_url+"microsoft/Phi-3.5-mini-instruct", | |
| # "Gemma-2B":base_url+"google/gemma-2b-it", | |
| # "Zephyr-7B-β":base_url+"HuggingFaceH4/zephyr-7b-beta", | |
| # "Llama-2":"meta-llama/Llama-2-7b-chat-hf" | |
| } | |
| #Pull info about the model to display | |
| model_info ={ | |
| "Mistral-7B": | |
| {'description':"""The Mistral model is able to have question and answer interactions.\n \ | |
| \nIt was created by the [**Mistral AI**](https://mistral.ai/news/announcing-mistral-7b/) team as has over **7 billion parameters.** \n""",}, | |
| "Phi-3.5": | |
| {'description':"""Phi-3.5-mini is a lightweight, state-of-the-art open model built upon datasets used for Phi-3 - synthetic data and filtered publicly available websites - with a focus on very high-quality, reasoning dense data. The model belongs to the Phi-3 model family and supports 128K token context length. The model underwent a rigorous enhancement process, incorporating both supervised fine-tuning, proximal policy optimization, and direct preference optimization to ensure precise instruction adherence and robust safety measures.""",} | |
| # "Gemma-7B": | |
| # {'description':"""The Gemma model is a **Large Language Model (LLM)** that's able to have question and answer interactions.\n \ | |
| # \nIt was created by the [**Google's AI Team**](https://blog.google/technology/developers/gemma-open-models/) team as has over **7 billion parameters.** \n""", | |
| # 'logo':'https://pbs.twimg.com/media/GG3sJg7X0AEaNIq.jpg'}, | |
| # "Gemma-2B": | |
| # {'description':"""The Gemma model is a **Large Language Model (LLM)** that's able to have question and answer interactions.\n \ | |
| # \nIt was created by the [**Google's AI Team**](https://blog.google/technology/developers/gemma-open-models/) team as has over **2 billion parameters.** \n""", | |
| # 'logo':'https://pbs.twimg.com/media/GG3sJg7X0AEaNIq.jpg'}, | |
| # "Zephyr-7B": | |
| # {'description':"""The Zephyr model is a **Large Language Model (LLM)** that's able to have question and answer interactions.\n \ | |
| # \nFrom Huggingface: \n\ | |
| # Zephyr is a series of language models that are trained to act as helpful assistants. \ | |
| # [Zephyr 7B Gemma](https://huggingface.co/HuggingFaceH4/zephyr-7b-gemma-v0.1)\ | |
| # is the third model in the series, and is a fine-tuned version of google/gemma-7b \ | |
| # that was trained on on a mix of publicly available, synthetic datasets using Direct Preference Optimization (DPO)\n""", | |
| # 'logo':'https://huggingface.co/HuggingFaceH4/zephyr-7b-gemma-v0.1/resolve/main/thumbnail.png'}, | |
| # "Zephyr-7B-β": | |
| # {'description':"""The Zephyr model is a **Large Language Model (LLM)** that's able to have question and answer interactions.\n \ | |
| # \nFrom Huggingface: \n\ | |
| # Zephyr is a series of language models that are trained to act as helpful assistants. \ | |
| # [Zephyr-7B-β](https://huggingface.co/HuggingFaceH4/zephyr-7b-beta)\ | |
| # is the second model in the series, and is a fine-tuned version of mistralai/Mistral-7B-v0.1 \ | |
| # that was trained on on a mix of publicly available, synthetic datasets using Direct Preference Optimization (DPO)\n""", | |
| # 'logo':'https://huggingface.co/HuggingFaceH4/zephyr-7b-alpha/resolve/main/thumbnail.png'}, | |
| } | |
| def format_promt(message, custom_instructions=None): | |
| prompt = "" | |
| if custom_instructions: | |
| prompt += f"[INST] {custom_instructions} [/INST]" | |
| prompt += f"[INST] {message} [/INST]" | |
| return prompt | |
| def reset_conversation(): | |
| ''' | |
| Resets Conversation | |
| ''' | |
| st.session_state.conversation = [] | |
| st.session_state.messages = [] | |
| return None | |
| models =[key for key in model_links.keys()] | |
| # Create the sidebar with the dropdown for model selection | |
| selected_model = st.sidebar.selectbox("Select Model", models) | |
| #Create a temperature slider | |
| temp_values = st.sidebar.slider('Select a temperature value', 0.0, 1.0, (0.5)) | |
| #Add reset button to clear conversation | |
| st.sidebar.button('Reset Chat', on_click=reset_conversation) #Reset button | |
| # Create model description | |
| st.sidebar.write(f"You're now chatting with **{selected_model}**") | |
| st.sidebar.markdown(model_info[selected_model]['description']) | |
| # st.sidebar.image(model_info[selected_model]['logo']) | |
| st.sidebar.markdown("*Generated content may be inaccurate or false.*") | |
| st.sidebar.markdown("\nLearn how to build this chatbot by original author of this chatbot [here](https://ngebodh.github.io/projects/2024-03-05/).") | |
| if "prev_option" not in st.session_state: | |
| st.session_state.prev_option = selected_model | |
| if st.session_state.prev_option != selected_model: | |
| st.session_state.messages = [] | |
| # st.write(f"Changed to {selected_model}") | |
| st.session_state.prev_option = selected_model | |
| reset_conversation() | |
| #Pull in the model we want to use | |
| repo_id = model_links[selected_model] | |
| st.subheader(f'AI - {selected_model}') | |
| # st.title(f'ChatBot Using {selected_model}') | |
| # Initialize chat history | |
| if "messages" not in st.session_state: | |
| st.session_state.messages = [] | |
| # Display chat messages from history on app rerun | |
| for message in st.session_state.messages: | |
| with st.chat_message(message["role"]): | |
| st.markdown(message["content"]) | |
| # Accept user input | |
| if prompt := st.chat_input(f"Hi I'm {selected_model}, ask me a question"): | |
| custom_instruction = "Act like a Human in conversation" | |
| # Display user message in chat message container | |
| with st.chat_message("user"): | |
| st.markdown(prompt) | |
| # Add user message to chat history | |
| st.session_state.messages.append({"role": "user", "content": prompt}) | |
| formated_text = format_promt(prompt, custom_instruction) | |
| # Display assistant response in chat message container | |
| with st.chat_message("assistant"): | |
| client = InferenceClient( | |
| model=model_links[selected_model],) | |
| # headers=headers) | |
| output = client.text_generation( | |
| formated_text, | |
| temperature=temp_values,#0.5 | |
| max_new_tokens=3000, | |
| stream=True | |
| ) | |
| response = st.write_stream(output) | |
| st.session_state.messages.append({"role": "assistant", "content": response}) | |