[TVMC] Separate model loading from model compilation in TVMC. (apache…

…#7739)

* add to init files for clean tvmc python

* adjust tests to new imports

* add to compiler.py

* update so model loads in drive_compile

* update test_compiler.py to load outside of tvmc.compile, need to correct one error

* fix mock.patch test

* remove merge artifact (circular import issue)

* change typo and merge artifact

* fix import in test_compiler.py

* black needed files

* remove unnecessary argument model_format from compile_module

* load before compile in conftest.py

* fix conftest.py issue

* fix typo in test_compiler.py

Co-authored-by: Jocelyn <jocelyn@pop-os.localdomain>
Co-authored-by: Josh Fromm <jwfromm@uw.edu>

python/tvm/driver/tvmc/compiler.py

@@ -111,16 +111,16 @@ def drive_compile(args):
         Zero if successfully completed
 
     """
+    mod, params = frontends.load_model(args.FILE, args.model_format, args.input_shapes)
 
     graph, lib, params, dumps = compile_model(
-        args.FILE,
+        mod,
+        params,
         args.target,
         args.dump_code,
         None,
-        args.model_format,
         args.tuning_records,
         args.desired_layout,
-        args.input_shapes,
     )
 
     if dumps:
@@ -131,14 +131,13 @@ def drive_compile(args):
 
 
 def compile_model(
-    path,
+    mod,
+    params,
     target,
     dump_code=None,
     target_host=None,
-    model_format=None,
     tuning_records=None,
     alter_layout=None,
-    shape_dict=None,
 ):
     """Compile a model from a supported framework into a TVM module.
 
@@ -148,8 +147,10 @@ def compile_model(
 
     Parameters
     ----------
-    path: str
-        Path to a file
+    mod: IRModule
+        The relay module to be compiled.
+    params: dict
+        A dictionary containing the module's parameters.
     target : str
         The target for which to compile. Can be a plain string or
         a path.
@@ -159,18 +160,13 @@ def compile_model(
     target_host : str, optional
         The target of the host machine if host-side code
         needs to be generated.
-    model_format: str, optional
-        A string representing a name of a frontend to be used
     tuning_records: str, optional
         Path to the file produced by the tuning to be used during
         compilation.
     alter_layout: str, optional
         The layout to convert the graph to. Note, the convert layout
         pass doesn't currently guarantee the whole of the graph will
         be converted to the chosen layout.
-    shape_dict: dict, optional
-        A mapping from input names to their shape. When present,
-        the default shapes in the model will be overwritten.
 
     Returns
     -------
@@ -185,7 +181,6 @@ def compile_model(
 
     """
     dump_code = [x.strip() for x in dump_code.split(",")] if dump_code else None
-    mod, params = frontends.load_model(path, model_format, shape_dict)
     config = {}
 
     if alter_layout:

tests/python/driver/tvmc/conftest.py

@@ -51,7 +51,8 @@ def get_sample_compiled_module(target_dir):
         temp_dir=target_dir,
     )
 
-    return tvmc.compiler.compile_model(model_file, target="llvm")
+    mod, params = tvmc.frontends.load_model(model_file)
+    return tvmc.compiler.compile_model(mod, params, target="llvm")
 
 
 # PyTest fixtures

tests/python/driver/tvmc/test_compiler.py

@@ -44,9 +44,9 @@ def test_save_dumps(tmpdir_factory):
 
 def verify_compile_tflite_module(model, shape_dict=None):
     pytest.importorskip("tflite")
-
+    mod, params = tvmc.load(model, shape_dict=shape_dict)
     graph, lib, params, dumps = tvmc.compile(
-        model, target="llvm", dump_code="ll", alter_layout="NCHW", shape_dict=shape_dict
+        mod, params, target="llvm", dump_code="ll", alter_layout="NCHW"
     )
 
     # check for output types
@@ -74,8 +74,10 @@ def test_compile_tflite_module(tflite_mobilenet_v1_1_quant):
 def test_cross_compile_aarch64_tflite_module(tflite_mobilenet_v1_1_quant):
     pytest.importorskip("tflite")
 
+    mod, params = tvmc.load(tflite_mobilenet_v1_1_quant)
     graph, lib, params, dumps = tvmc.compile(
-        tflite_mobilenet_v1_1_quant,
+        mod,
+        params,
         target="llvm -device=arm_cpu -mtriple=aarch64-linux-gnu -mattr='+neon'",
         dump_code="asm",
     )
@@ -91,7 +93,8 @@ def test_compile_keras__save_module(keras_resnet50, tmpdir_factory):
     # some CI environments wont offer tensorflow/Keras, so skip in case it is not present
     pytest.importorskip("tensorflow")
 
-    graph, lib, params, dumps = tvmc.compile(keras_resnet50, target="llvm", dump_code="ll")
+    mod, params = tvmc.load(keras_resnet50)
+    graph, lib, params, dumps = tvmc.compile(mod, params, target="llvm", dump_code="ll")
 
     expected_temp_dir = tmpdir_factory.mktemp("saved_output")
     expected_file_name = "saved.tar"
@@ -109,8 +112,10 @@ def test_cross_compile_aarch64_keras_module(keras_resnet50):
     # some CI environments wont offer tensorflow/Keras, so skip in case it is not present
     pytest.importorskip("tensorflow")
 
+    mod, params = tvmc.load(keras_resnet50)
     graph, lib, params, dumps = tvmc.compile(
-        keras_resnet50,
+        mod,
+        params,
         target="llvm -device=arm_cpu -mtriple=aarch64-linux-gnu -mattr='+neon'",
         dump_code="asm",
     )
@@ -126,10 +131,8 @@ def test_cross_compile_aarch64_keras_module(keras_resnet50):
 def verify_compile_onnx_module(model, shape_dict=None):
     # some CI environments wont offer onnx, so skip in case it is not present
     pytest.importorskip("onnx")
-
-    graph, lib, params, dumps = tvmc.compile(
-        model, target="llvm", dump_code="ll", shape_dict=shape_dict
-    )
+    mod, params = tvmc.load(model, shape_dict=shape_dict)
+    graph, lib, params, dumps = tvmc.compile(mod, params, target="llvm", dump_code="ll")
 
     # check for output types
     assert type(graph) is str
@@ -156,8 +159,10 @@ def test_cross_compile_aarch64_onnx_module(onnx_resnet50):
     # some CI environments wont offer onnx, so skip in case it is not present
     pytest.importorskip("onnx")
 
+    mod, params = tvmc.load(onnx_resnet50)
     graph, lib, params, dumps = tvmc.compile(
-        onnx_resnet50,
+        mod,
+        params,
         target="llvm -device=arm_cpu -mtriple=aarch64-linux-gnu -mattr=+neon",
         dump_code="asm",
     )
@@ -173,9 +178,10 @@ def test_cross_compile_aarch64_onnx_module(onnx_resnet50):
 @tvm.testing.requires_opencl
 def test_compile_opencl(tflite_mobilenet_v1_0_25_128):
     pytest.importorskip("tflite")
-
+    mod, params = tvmc.load(tflite_mobilenet_v1_0_25_128)
     graph, lib, params, dumps = tvmc.compile(
-        tflite_mobilenet_v1_0_25_128,
+        mod,
+        params,
         target="opencl --host=llvm",
         alter_layout="NCHW",
     )
@@ -193,9 +199,9 @@ def test_compile_opencl(tflite_mobilenet_v1_0_25_128):
 )
 def test_compile_tflite_module_with_external_codegen(tflite_mobilenet_v1_1_quant):
     pytest.importorskip("tflite")
-
+    mod, params = tvmc.load(tflite_mobilenet_v1_1_quant)
     graph, lib, params, dumps = tvmc.compile(
-        tflite_mobilenet_v1_1_quant, target="ethos-n77, llvm", dump_code="relay"
+        mod, params, target="ethos-n77, llvm", dump_code="relay"
     )
 
     # check for output types
@@ -207,7 +213,7 @@ def test_compile_tflite_module_with_external_codegen(tflite_mobilenet_v1_1_quant
 
 @mock.patch("tvm.relay.build")
 @mock.patch("tvm.driver.tvmc.composite_target.get_codegen_by_target")
-@mock.patch("tvm.driver.tvmc.frontends.load_model")
+@mock.patch("tvm.driver.tvmc.load")
 @mock.patch("tvm.transform.PassContext")
 def test_compile_check_configs_composite_target(mock_pc, mock_fe, mock_ct, mock_relay):
     mock_codegen = {}
@@ -218,9 +224,8 @@ def test_compile_check_configs_composite_target(mock_pc, mock_fe, mock_ct, mock_
     mock_ct.return_value = mock_codegen
     mock_relay.return_value = mock.MagicMock()
 
-    graph, lib, params, dumps = tvmc.compile(
-        "no_file_needed", target="mockcodegen -testopt=value, llvm"
-    )
+    mod, params = tvmc.load("no_file_needed")
+    graph, lib, params, dumps = tvmc.compile(mod, params, target="mockcodegen -testopt=value, llvm")
 
     mock_pc.assert_called_once_with(
         opt_level=3, config={"relay.ext.mock.options": {"testopt": "value"}}