Create app.py

app.py

@@ -0,0 +1,41 @@
+from peft import AutoPeftModelForCausalLM
+from transformers import AutoTokenizer, pipeline
+import gradio as gr
+
+# Load the model
+model = AutoPeftModelForCausalLM.from_pretrained("Moritz-Pfeifer/financial-times-classification-llama-2-7b-v1.3")
+tokenizer = AutoTokenizer.from_pretrained("Moritz-Pfeifer/financial-times-classification-llama-2-7b-v1.3")
+
+def predict_text(test, model, tokenizer):
+    prompt = f"""
+            You are given a news article regarding the greater Boston area.
+            Analyze the sentiment of the article enclosed in square brackets,
+            determine if it is positive, negative or other, and return the answer as the corresponding sentiment label
+            "positive" or "negative". If the sentiment is neither positive or negative, return "other".
+
+            [{test}] ="""
+    pipe = pipeline(task="text-generation",
+                        model=model,
+                        tokenizer=tokenizer,
+                        max_new_tokens = 1,
+                        temperature = 0.1,
+                       )
+    result = pipe(prompt)
+    answer = result[0]['generated_text'].split("=")[-1]
+    # print(answer)
+    if "positive" in answer.lower():
+        return "positive"
+    elif "negative" in answer.lower():
+        return "negative"
+    else:
+        return "other"
+
+def predict(input_text):
+  return predict_text(input_text, model, tokenizer)
+
+
+interface = gr.Interface(fn=predict, inputs="text", outputs="text", title="Text Classifier", description="Insert your text and get the classification result.")
+interface.launch()
+
+if __name__ == "__main__":
+    interface.launch(share=True)