Spaces:
Runtime error
Runtime error
| import together | |
| import os | |
| # set your API key | |
| together.api_key = "c9909567768fbf1a69fbd94c758e432f0a05a6755c32dced992ac6640a8cfd79" | |
| # list available models and descriptons | |
| models = together.Models.list() | |
| together.Models.start("togethercomputer/llama-2-7b-chat") | |
| from langchain.llms import Together | |
| llm = Together( | |
| model="togethercomputer/llama-2-7b-chat", | |
| temperature=0.7, | |
| max_tokens=128, | |
| top_k=1, | |
| together_api_key="c9909567768fbf1a69fbd94c758e432f0a05a6755c32dced992ac6640a8cfd79" | |
| ) | |
| from langchain.vectorstores import Chroma | |
| from langchain.text_splitter import RecursiveCharacterTextSplitter | |
| from langchain.chains import RetrievalQA | |
| from langchain.document_loaders import TextLoader | |
| from langchain.document_loaders import PyPDFLoader | |
| from langchain.document_loaders import DirectoryLoader | |
| pdf_path = os.path.join(os.path.dirname(__file__), 'Production-Table - Sheet1 (2).pdf') | |
| loader = PyPDFLoader(pdf_path) | |
| documents = loader.load() | |
| #splitting the text into | |
| text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200) | |
| texts = text_splitter.split_documents(documents) | |
| from langchain.embeddings import HuggingFaceBgeEmbeddings | |
| model_name = "BAAI/bge-base-en" | |
| encode_kwargs = {'normalize_embeddings': True} # set True to compute cosine similarity | |
| model_norm = HuggingFaceBgeEmbeddings( | |
| model_name=model_name, | |
| model_kwargs={'device': 'cuda'}, | |
| encode_kwargs=encode_kwargs | |
| ) | |
| # Embed and store the texts | |
| # Supplying a persist_directory will store the embeddings on disk | |
| persist_directory = 'db' | |
| ## Here is the nmew embeddings being used | |
| embedding = model_norm | |
| vectordb = Chroma.from_documents(documents=texts, | |
| embedding=embedding, | |
| persist_directory=persist_directory) | |
| retriever = vectordb.as_retriever(search_kwargs={"k": 5}) | |
| ## Default LLaMA-2 prompt style | |
| B_INST, E_INST = "[INST]", "[/INST]" | |
| B_SYS, E_SYS = "<<SYS>>\n", "\n<</SYS>>\n\n" | |
| DEFAULT_SYSTEM_PROMPT = """\ | |
| You are a helpful, respectful and honest assistant of a production company. You should honestly answer the user's query using the knowledge of the company's production documents uploaded. | |
| If a question does not make any sense, or is not factually coherent, explain why instead of answering something not correct. If you don't know the answer to a question, please don't share false information.""" | |
| def get_prompt(instruction, new_system_prompt=DEFAULT_SYSTEM_PROMPT ): | |
| SYSTEM_PROMPT = B_SYS + new_system_prompt + E_SYS | |
| prompt_template = B_INST + SYSTEM_PROMPT + instruction + E_INST | |
| return prompt_template | |
| sys_prompt = """You are a helpful, respectful and honest assistant of a production company. You should honestly answer the user's query using the knowledge of the company's production documents uploaded. | |
| If a question does not make any sense, or is not factually coherent, explain why instead of answering something not correct. If you don't know the answer to a question, please don't share false information.""" | |
| instruction = """CONTEXT:/n/n {context}/n | |
| Question: {question}""" | |
| get_prompt(instruction, sys_prompt) | |
| from langchain.prompts import PromptTemplate | |
| prompt_template = get_prompt(instruction, sys_prompt) | |
| llama_prompt = PromptTemplate( | |
| template=prompt_template, input_variables=["context", "question"] | |
| ) | |
| chain_type_kwargs = {"prompt": llama_prompt} | |
| from langchain.schema import prompt | |
| # create the chain to answer questions | |
| qa_chain = RetrievalQA.from_chain_type(llm=llm, | |
| chain_type="stuff", | |
| retriever=retriever, | |
| chain_type_kwargs=chain_type_kwargs, | |
| return_source_documents=True) | |
| ## Cite sources | |
| import textwrap | |
| def wrap_text_preserve_newlines(text, width=110): | |
| # Split the input text into lines based on newline characters | |
| lines = text.split('\n') | |
| # Wrap each line individually | |
| wrapped_lines = [textwrap.fill(line, width=width) for line in lines] | |
| # Join the wrapped lines back together using newline characters | |
| wrapped_text = '\n'.join(wrapped_lines) | |
| return wrapped_text | |
| def process_llm_response(llm_response): | |
| print(wrap_text_preserve_newlines(llm_response['result'])) | |
| print('\n\nSources:') | |
| for source in llm_response["source_documents"]: | |
| print(source.metadata['source']) | |
| import gradio as gr | |
| with gr.Blocks() as demo: | |
| chatbot = gr.Chatbot() | |
| msg = gr.Textbox() | |
| clear = gr.Button("Clear") | |
| def user(user_message, history): | |
| return "", history + [[user_message, None]] | |
| def bot(history): | |
| print("Question: ", history[-1][0]) | |
| #wrap_text_preserve_newlines(llm_response['result']) | |
| #bot_message = process_llm_response(qa_chain(history[-1][0])) | |
| bot_message = wrap_text_preserve_newlines((qa_chain(history[-1][0]))['result']) | |
| print("Response: ", bot_message) | |
| history[-1][1] = "" | |
| history[-1][1] += bot_message | |
| return history | |
| msg.submit(user, [msg, chatbot], [msg, chatbot], queue=False).then(bot, chatbot, chatbot) | |
| clear.click(lambda: None, None, chatbot, queue=False) | |
| demo.queue() | |
| demo.launch(debug = True) |