Upload app.py
Browse files
app.py
ADDED
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
from transformers import pipeline
|
3 |
+
import numpy as np
|
4 |
+
|
5 |
+
hhem = pipeline("text-classification", model="vectara/hallucination_evaluation_model")
|
6 |
+
|
7 |
+
def get_hhem_score(sentence1, sentence2):
|
8 |
+
output = hhem(f'{sentence1} [SEP] {sentence2}')
|
9 |
+
score = np.round(output[0]['score'], 4)
|
10 |
+
|
11 |
+
return score
|
12 |
+
|
13 |
+
demo = gr.Interface(
|
14 |
+
fn=get_hhem_score,
|
15 |
+
inputs=[
|
16 |
+
gr.components.Textbox(label="Sentence 1"),
|
17 |
+
gr.components.Textbox(label="Sentence 2"),
|
18 |
+
],
|
19 |
+
outputs=gr.components.Label(num_top_classes=1, label='HHEM Score'),
|
20 |
+
examples=[
|
21 |
+
["Vectara provides RAG-as-a-service", "RAG-as-a-service is provided by Vectara"],
|
22 |
+
["The cat sat on the mat.", "A feline was resting on a small rug."],
|
23 |
+
["The quick brown fox jumps over the lazy dog.", "A fast red fox leaps across a sleepy canine."],
|
24 |
+
],
|
25 |
+
cache_examples=False,
|
26 |
+
allow_flagging="never",
|
27 |
+
flagging_options=None,
|
28 |
+
title="HHEM Demo",
|
29 |
+
description="This demo uses Vectara's Hallucination Evaluation model (HHEM) to calculate factual consistency between two input sentences.",
|
30 |
+
)
|
31 |
+
|
32 |
+
# Launch the demo
|
33 |
+
demo.launch()
|