Fix test

Author: jackzhxng

examples/models/llama/export_llama_lib.py

@@ -905,7 +905,7 @@ def _to_edge_and_lower_llama_xnnpack(
     return builder.to_executorch(passes=additional_passes)
 
 
-def _to_edge_and_lower_llama(
+def _to_edge_and_lower_llama(  # noqa: C901
     builder_exported,
     modelname,
     additional_passes,

examples/models/llava/export_llava.py

@@ -162,7 +162,7 @@ def forward(self, images):
             use_kv_cache=True,
             example_inputs=(resized,),
             dynamic_shapes=dynamic_shapes,
-            args=None,
+            config=None,
         )
         .export()
         .pt2e_quantize([quantizer])

extension/llm/export/builder.py

@@ -204,7 +204,7 @@ def _export(self, module: Optional[torch.nn.Module] = None) -> ExportedProgram:
         # 1. torch.nn.attention.sdpa_kernel([SDPBackend.MATH]) is for bypassing the dynamo error when tracing
         # 2. torch.no_grad() is for getting rid of the dropout (not sure why training ops will show up)
         with torch.nn.attention.sdpa_kernel([SDPBackend.MATH]), torch.no_grad():
-            if self.config.backend.qnn.enabled:
+            if self.config and self.config.backend.qnn.enabled:
                 # TODO: this is temporary, as qnn flow does not work with new, non-functional export IR.
                 # See issue: https://github.com/pytorch/executorch/issues/7373
 
@@ -250,7 +250,7 @@ def export(self) -> "LLMEdgeManager":
         # Persisting those changes back to an ExportedProgram will require
         # an additional export().
         self.pre_autograd_graph_module = exported_module.module()
-        if self.config.export.export_only:
+        if self.config and self.config.export.export_only:
             torch.export.save(exported_module, self.config.export.output_name)
         return self
 
@@ -415,7 +415,7 @@ def export_to_edge(self) -> "LLMEdgeManager":
                 self.export()
 
             override_export_behaviour = contextlib.nullcontext()
-            if self.config.backend.qnn.enabled:
+            if self.config and self.config.backend.qnn.enabled:
                 override_export_behaviour = patch.object(
                     torch._utils_internal,
                     "export_training_ir_rollout_check",