disanle batching

generate.py

@@ -26,11 +26,11 @@ logger.warning("Loading model...")
 if torch.backends.mps.is_available():
     device = "mps"
     model_id = "Qwen/Qwen1.5-0.5B-Chat"
-    batch_size = 4
+    batch_size = 1  # batching generates duplicates
 else:
     device = "cuda"
     model_id = "google/gemma-2b-it"
-    batch_size = 4
+    batch_size = 1  # batching generates duplicates
 
 model = models.transformers(model_id, device=device)