File size: 1,855 Bytes
b686079 | 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 | import os
import re
import gradio as gr
from transformers import pipeline
pipe = pipeline(
"text-generation",
model="ProtectDemocracy/threat-detector",
token=os.getenv("HF_TOKEN"),
)
THREAT_PROMPT: str = """Analyze the following social media post from the user and determine whether or not a threat of violence is present or not. Note that posts that are discussing the news, or simply responding to other threatening posts, are not themselves threatening. Respond with a 0 if there is no violent threat, or a 1 if there is a violent threat."""
# Define the function the interface will use
def generate_text(prompt):
messages = [
{"role": "system", "content": THREAT_PROMPT},
{"role": "user", "content": prompt},
]
full_prompt = pipe.tokenizer.apply_chat_template(
messages, tokenize=False, add_generation_prompt=True
)
# Generate text, restricting the number of new tokens
output = pipe(full_prompt, max_new_tokens=1, do_sample=False)
generated_text = output[0]["generated_text"]
response = generated_text[len(full_prompt) :]
return response.strip()
# Create the Gradio Interface
iface = gr.Interface(
fn=generate_text,
inputs=gr.Textbox(
lines=5,
label="Input",
placeholder="Enter a description of a potential threat...",
),
outputs=gr.Textbox(label="Threat Level (0 or 1)"),
title="Threat Detector",
description="This tool analyzes a description of a potential threat and classifies it as threatening (1) or nonthreatening (0)",
examples=[
["Just had the most delicious lunch :) #TacoTuesday"],
["I am going to commit an imminent act of unlawful violence!"],
[
"Really wish he had hit that shot! Now the Lakers are probably going to lose."
],
],
)
# Launch
iface.launch()
|