cleanup

machine-learning/app/models/base.py

@@ -9,7 +9,7 @@
 import onnx
 import onnxruntime as ort
 from huggingface_hub import snapshot_download
-from onnx.shape_inference import infer_shapes
+from onnx.shape_inference import infer_shapes_path
 from onnx.tools.update_model_dims import update_inputs_outputs_dims
 from typing_extensions import Buffer
 import ann.ann
@@ -117,8 +117,7 @@ def _make_session(self, model_path: Path) -> AnnSession | ort.InferenceSession:
             model_path = onnx_path
 
         if any(provider in STATIC_INPUT_PROVIDERS for provider in self.providers):
-            static_path = model_path.parent / "static_1" / "model.onnx"
-            static_path.parent.mkdir(parents=True, exist_ok=True)
+            static_path = model_path.parent / "model_static_1.onnx"
             if not static_path.is_file():
                 self._convert_to_static(model_path, static_path)
             model_path = static_path
@@ -138,29 +137,24 @@ def _make_session(self, model_path: Path) -> AnnSession | ort.InferenceSession:
         return session
 
     def _convert_to_static(self, source_path: Path, target_path: Path) -> None:
-        inferred = infer_shapes(onnx.load(source_path))
-        inputs = self._get_static_dims(inferred.graph.input)
-        outputs = self._get_static_dims(inferred.graph.output)
+        infer_shapes_path(source_path, strict_mode=True)
+        proto = onnx.load(source_path, load_external_data=False)
+        inputs = self._get_static_dims(proto.graph.input)
+        outputs = self._get_static_dims(proto.graph.output)
 
-        # check_model gets called in update_inputs_outputs_dims and doesn't work for large models
+        # check_model gets called in update_inputs_outputs_dims
         check_model = onnx.checker.check_model
         try:
-
+            
             def check_model_stub(*args: Any, **kwargs: Any) -> None:
                 pass
 
             onnx.checker.check_model = check_model_stub
-            updated_model = update_inputs_outputs_dims(inferred, inputs, outputs)
+            updated_model = update_inputs_outputs_dims(proto, inputs, outputs)
         finally:
             onnx.checker.check_model = check_model
 
-        onnx.save(
-            updated_model,
-            target_path,
-            save_as_external_data=True,
-            all_tensors_to_one_file=False,
-            size_threshold=1048576,
-        )
+        onnx.save(updated_model, target_path)
 
     def _get_static_dims(self, graph_io: Any, dim_size: int = 1) -> dict[str, list[int]]:
         return {