|
from huggingface_hub import InferenceClient |
|
import gradio as gr |
|
from langchain.embeddings import HuggingFaceEmbeddings |
|
from langchain.vectorstores import FAISS |
|
|
|
embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2") |
|
vector_store = FAISS.load_local("TaoGPT-Embeddings", embeddings) |
|
|
|
client = InferenceClient( |
|
"mistralai/Mistral-7B-Instruct-v0.1" |
|
) |
|
|
|
NOMIC = """ |
|
<!DOCTYPE html> |
|
<html> |
|
<head> |
|
<title>TaoGPT - DataMap</title> |
|
<style> |
|
iframe { |
|
width: 100%; |
|
height: 600px; /* You can adjust the height as needed */ |
|
border: 0; |
|
} |
|
</style> |
|
</head> |
|
<body> |
|
<iframe |
|
src="https://atlas.nomic.ai/map/a59e104a-e4d0-4701-bbd9-3d552aae92a2/a3a2aacd-0787-4389-863a-ab3670015367?xs=-41.70341&xf=41.36850&ys=-23.57587&yf=23.20673" |
|
></iframe> |
|
</body> |
|
</html> |
|
""" |
|
RAG = True |
|
|
|
def format_prompt(message, history): |
|
global RAG |
|
if RAG == True: |
|
results = vector_store.similarity_search(message ,k=3) |
|
context = [result.page_content for result in results] |
|
context = "\n\n".join(context) |
|
print(context) |
|
prompt = "<s>" |
|
for user_prompt, bot_response in history: |
|
prompt += f"[INST] {user_prompt} [/INST]" |
|
prompt += f" {bot_response}</s> " |
|
prompt += f"[INST]Given the following Information:\n{context} \n answer the following question {message} [/INST]" |
|
return prompt |
|
else: |
|
prompt = "<s>" |
|
for user_prompt, bot_response in history: |
|
prompt += f"[INST] {user_prompt} [/INST]" |
|
prompt += f" {bot_response}</s> " |
|
prompt += f"[INST] {message} [/INST]" |
|
return prompt |
|
|
|
|
|
def generate( |
|
prompt, history, temperature=0.9, max_new_tokens=256, top_p=0.95, repetition_penalty=1.0, |
|
): |
|
temperature = float(temperature) |
|
if temperature < 1e-2: |
|
temperature = 1e-2 |
|
top_p = float(top_p) |
|
|
|
generate_kwargs = dict( |
|
temperature=temperature, |
|
max_new_tokens=max_new_tokens, |
|
top_p=top_p, |
|
repetition_penalty=repetition_penalty, |
|
do_sample=True, |
|
seed=42, |
|
) |
|
|
|
formatted_prompt = format_prompt(prompt, history) |
|
|
|
stream = client.text_generation(formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=False) |
|
output = "" |
|
|
|
for response in stream: |
|
output += response.token.text |
|
yield output |
|
return output |
|
|
|
|
|
additional_inputs=[ |
|
gr.Slider( |
|
label="Temperature", |
|
value=0.9, |
|
minimum=0.0, |
|
maximum=1.0, |
|
step=0.05, |
|
interactive=True, |
|
info="Higher values produce more diverse outputs", |
|
), |
|
gr.Slider( |
|
label="Max new tokens", |
|
value=256, |
|
minimum=0, |
|
maximum=1048, |
|
step=64, |
|
interactive=True, |
|
info="The maximum numbers of new tokens", |
|
), |
|
gr.Slider( |
|
label="Top-p (nucleus sampling)", |
|
value=0.90, |
|
minimum=0.0, |
|
maximum=1, |
|
step=0.05, |
|
interactive=True, |
|
info="Higher values sample more low-probability tokens", |
|
), |
|
gr.Slider( |
|
label="Repetition penalty", |
|
value=1.2, |
|
minimum=1.0, |
|
maximum=2.0, |
|
step=0.05, |
|
interactive=True, |
|
info="Penalize repeated tokens", |
|
) |
|
] |
|
|
|
css = """ |
|
#mkd { |
|
height: 500px; |
|
overflow: auto; |
|
border: 1px solid #ccc; |
|
} |
|
""" |
|
|
|
|
|
|
|
with gr.Blocks(css=css) as demo: |
|
|
|
|
|
|
|
gr.HTML("<h1><center>TaoGPT<center></h1>") |
|
gr.HTML("<h3><center>TaoGPT is Fine-tuned Mistal-7B model on TaoScience related Information Check out- <a href='https://github.com/agencyxr/taogpt7B'>Github Repo</a> For More Information. 💬<h3><center>") |
|
with gr.Row(): |
|
with gr.Column(): |
|
gr.HTML("<h3>Chat with TaoGPT</h3>") |
|
gr.ChatInterface( |
|
generate, |
|
additional_inputs=additional_inputs, |
|
examples=[["What is TaoScience"], ["Give me a Summary about TaoScience"]] |
|
) |
|
RAG_Checkbox = gr.Checkbox(label="Use Retrival Augmented Generation" , value=True , interactive=False) |
|
with gr.Column(): |
|
gr.HTML("<h3>Look into the Dataset we used to Finetune our Model</h3>") |
|
gr.HTML(NOMIC) |
|
|
|
demo.queue(concurrency_count=75, max_size=100).launch(debug=True) |
