Spaces:
Running
Running
File size: 3,535 Bytes
424ded8 845c72e 0439b4c dae9913 0439b4c 424ded8 f93e307 11954b4 c6e7871 11954b4 c6e7871 11954b4 424ded8 061103a 48a339d 424ded8 c6e7871 469ea1a 424ded8 c6e7871 424ded8 c6e7871 |
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 |
import gradio as gr
from transformers import pipeline, AutoTokenizer
import os
# Retrieve the Hugging Face API token from environment variables
hf_token = os.getenv("HF_TOKEN")
if not hf_token:
raise ValueError("API token is not set. Please set the HF_TOKEN environment variable in Space Settings.")
def authenticate_and_generate(message, history, system_message, max_tokens, temperature, top_p):
try:
# Initialize the text-generation pipeline with the provided token
text_generator = pipeline("text-generation", model="microsoft/Phi-3-mini-4k-instruct", use_auth_token=hf_token, trust_remote_code=True)
# Load the tokenizer separately if needed
try:
tokenizer = AutoTokenizer.from_pretrained("microsoft/Phi-3-mini-4k-instruct", use_auth_token=hf_token)
except Exception as e:
raise RuntimeError(f"Failed to load the tokenizer: {str(e)}")
if text_generator.tokenizer is None:
raise RuntimeError("The tokenizer is not available. Check the model and API token.")
# Ensure that system_message is a string
system_message = str(system_message)
# Construct the prompt with system message, history, and user input
history_str = "\n".join([f"User: {str(msg[0])}\nAssistant: {str(msg[1])}" for msg in history if isinstance(msg, (tuple, list)) and len(msg) == 2])
prompt = system_message + "\n" + history_str
prompt += f"\nUser: {message}\nAssistant:"
# Generate a response using the model
response = text_generator(prompt, max_length=max_tokens, temperature=temperature, top_p=top_p, do_sample=True, truncation=True)
# Extract the generated text from the response list
assistant_response = response[0]['generated_text']
# Optionally trim the assistant response if it includes the prompt again
assistant_response = assistant_response.split("Assistant:", 1)[-1].strip()
return assistant_response
except Exception as e:
return str(e) # Return the error message for debugging
athena = gr.ChatInterface(
fn=authenticate_and_generate,
additional_inputs=[
gr.Textbox(
value="""
You are a marketing-minded content writer for Plan.com (a UK telecommunications company).
You will be provided a bullet-point list of guidelines from which to generate an article to be published in the company News section of the website.
Please follow these guidelines:
- Always speak using British English expressions, syntax, and spelling.
- Make the articles engaging and fun, but also professional and informative.
To provide relevant contextual information about the company, please source information from the following websites:
- https://plan.com/our-story
- https://plan.com/products-services
- https://plan.com/features/productivity-and-performance
- https://plan.com/features/security-and-connectivity
- https://plan.com/features/connectivity-and-cost
""",
label="System message"
),
gr.Slider(minimum=1, maximum=4096, value=512, step=1, label="Max new tokens"),
gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p (nucleus sampling)"),
],
)
if __name__ == "__main__":
athena.launch()
|