Update app.py

app.py

@@ -79,6 +79,7 @@ def load_pipeline(model_name):
     Load and cache a transformers pipeline for text generation.
     Tries bfloat16, falls back to float16 or float32 if unsupported.
     """
+    global PIPELINES
     
     if model_name in PIPELINES.keys():
         return PIPELINES[model_name]
@@ -98,7 +99,6 @@ def load_pipeline(model_name):
                 device_map=device,
                 trust_remote_code=True,
             )
-    global PIPELINES
     PIPELINES[model_name] = {"tokenizer": tokenizer, "model": model}
     return {"tokenizer": tokenizer, "model": model}