Update app.py

app.py

@@ -66,7 +66,7 @@ def process_text(text, count):
             'model.embed_tokens')(inputs).to(device)
     prediction = text_tokenizer.batch_decode(
          tuned_phi2.generate(
-                inputs_embeds=inputs,
+                inputs_embeds=input_embeds,
                 max_new_tokens=30,
                 bos_token_id=text_tokenizer.bos_token_id,
                 eos_token_id=text_tokenizer.eos_token_id,