Upload 2 files

app.py

@@ -0,0 +1,39 @@
+from fastapi import FastAPI
+import gradio as gr
+
+from transformers import pipeline
+from gradio.components import Textbox
+
+app = FastAPI()
+
+# Load the sentiment analysis pipeline with DistilBERT
+distilbert_pipeline = pipeline("sentiment-analysis", model="distilbert-base-uncased-finetuned-sst-2-english")
+label_map = {"POSITIVE":"OTHER", "NEGATIVE":"SENSITIVE"}
+
+input1 = Textbox(lines=2, placeholder="Type your text here...")
+
+@app.get("/")
+async def root():
+    def predict_sentiment(text):
+        """
+        Predicts the sentiment of the input text using DistilBERT.
+        :param text: str, input text to analyze.
+        :return: str, predicted sentiment and confidence score.
+        """
+        result = distilbert_pipeline(text)[0]
+        label = label_map[result['label']]
+        score = result['score']
+        return f"TAG: {label}, Confidence: {score:.2f}"
+
+    # Create a Gradio interface
+    text_input = gr.Interface(fn=predict_sentiment,
+                         inputs=input1,
+                         outputs="text",
+                         title="Talk2Loop Sensitive statement tags",
+                         description="This model predicts the sensitivity of the input text. Enter a sentence to see if it's sensitive or not.")
+
+    return text_input.launch(share=True, host="0.0.0.0", port=8000)
+
+# Launch the interface
+app = gr.mount_gradio_app(app, text_input, path="/")
+# iface.launch(port=8000)

requirements.txt

@@ -0,0 +1,6 @@
+fastapi
+uvicorn
+gunicorn
+gradio
+torch
+transformers