Spaces:
Sleeping
Sleeping
File size: 2,990 Bytes
7416d8a 5c0e14a 7416d8a 5c0e14a 7416d8a 08eb742 5c0e14a 7416d8a 5c0e14a 7416d8a 5c0e14a 7416d8a 5c0e14a 7416d8a 17fba42 7416d8a 44b3bcb 5c0e14a 44b3bcb |
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 |
import json
from huggingface_hub import InferenceClient
import gradio as gr
import random
API_URL = "https://api-inference.huggingface.co/models/"
client = InferenceClient(
"mistralai/Mistral-7B-Instruct-v0.1"
)
def format_prompt(message, history):
prompt = "You're a helpful assistant."
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=2048, 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=random.randint(0, 10**7),
)
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
def load_database():
try:
# Attempt to load the database from JSON
with open("database.json", "r", encoding="utf-8") as f:
return json.load(f)
except (FileNotFoundError, json.JSONDecodeError):
# Handle potential errors gracefully
print("Error loading database: File not found or invalid format. Creating an empty database.")
return [] # Return an empty list if database loading fails
def save_database(data):
try:
# Save the updated database to JSON
with open("database.json", "w", encoding="utf-8") as f:
json.dump(data, f, indent=4)
except (IOError, json.JSONEncodeError):
# Handle potential errors gracefully
print("Error saving database: Encountered an issue while saving.")
def chat_interface(message):
database = load_database()
# Check if the question already exists in the database
if (message, None) not in database:
# If not, generate a response and add it to the database
response = generate(message, history=[])
database.append((message, response))
save_database(database)
else:
# If it does, retrieve the stored response
_, stored_response = next(item for item in database if item[0] == message)
response = stored_response
return response
with gr.Blocks(theme=gr.themes.Soft()) as demo:
# Use "Textbox" components for both input and output
input_textbox = gr.Textbox(label="Your question")
output_textbox = gr.Textbox(label="Assistant's response", value="", editable=False)
# Use demo.launch instead of demo.queue().launch()
demo.launch(fn=chat_interface, inputs=input_textbox, outputs=output_textbox)
|