Subgraph Shape propagation

graph_net/apply_sample_pass.py

@@ -39,23 +39,32 @@ def _get_handler(args):
 def main(args):
     handler = _get_handler(args)
     if args.model_path is not None:
+        assert not hasattr(handler, "BEGIN")
+        assert not hasattr(handler, "END")
         handle_model_path(handler, args.model_path)
     elif args.use_subprocess:
+        assert not hasattr(handler, "BEGIN")
+        assert not hasattr(handler, "END")
         handle_model_path_list_in_subprocess(args)
     else:
         handle_model_path_list_in_current_process(handler, args)
 
 
 def handle_model_path_list_in_current_process(handler, args):
-    for model_path in _get_model_path_list(args):
+    rel_model_paths = list(_get_model_path_list(args))
+    if hasattr(handler, "BEGIN"):
+        handler.BEGIN(rel_model_paths)
+    for rel_model_path in rel_model_paths:
         try:
-            handle_model_path(handler, model_path)
+            handle_model_path(handler, rel_model_path)
         except KeyboardInterrupt:
             print("KeyboardInterrupt")
             return
         except Exception:
             print("------------[apply_sample_pass failed]------------", flush=True)
             traceback.print_exc()
+    if hasattr(handler, "END"):
+        handler.END(rel_model_paths)
 
 
 def handle_model_path_list_in_subprocess(args):

graph_net/bash_templates/apply_sample_pass_sh.txt

@@ -0,0 +1,16 @@
+#!/bin/bash
+
+GRAPH_NET_ROOT=$(python3 -c "import graph_net; import os; print(os.path.dirname(os.path.dirname(graph_net.__file__)))")
+
+python3 -m graph_net.apply_sample_pass \
+    --model-path-list "customize_your_model_path_list" \
+    --sample-pass-file-path "$GRAPH_NET_ROOT/graph_net/customize_your_sample_pass.py" \
+    --sample-pass-class-name customize_your_class_name \
+    --sample-pass-config $(base64 -w 0 <<EOF
+{
+    "resume": true,
+    "model_path_prefix": "/customize_your_model_path_prefix",
+    "output_dir": "/customize_your_output_file"
+}
+EOF
+)

graph_net/sample_pass/group_ranges_from_subgraph_sources.py

@@ -0,0 +1,102 @@
+from graph_net.sample_pass.sample_pass import SamplePass
+from pathlib import Path
+import json
+
+
+class GroupRangesFromSubgraphSources(SamplePass):
+    def __init__(self, config=None):
+        super().__init__(config)
+        self.original_graph_rel_model_path2ranges: dict[str, list[(int, int)]] = {}
+        self.original_graph_rel_model_path2subgraph_rel_model_paths: dict[
+            str, list[str]
+        ] = {}
+
+    def declare_config(
+        self,
+        subgraph_model_path_prefix: str,
+        output_dir: str,
+        subgraph_sources_json_file_name: str = "subgraph_sources.json",
+        output_json_file_name: str = "grouped_ranges_from_subgraph_sources.json",
+        output_json_key: str = "grouped_ranges_from_subgraph_sources",
+        output_json_subgraph_rel_model_path_key: str = "subgraph_rel_model_paths",
+    ):
+        pass
+
+    def __call__(self, subgraph_rel_model_path: str):
+        model_path = (
+            Path(self.config["subgraph_model_path_prefix"])
+            / subgraph_rel_model_path
+            / self.config["subgraph_sources_json_file_name"]
+        )
+        subgraph_sources = json.load(open(model_path))
+        for original_graph_rel_model_path, subgraph_ranges in subgraph_sources.items():
+            self._collect_original_graph_rel_model_path2ranges(
+                original_graph_rel_model_path, subgraph_ranges
+            )
+            self._collect_original_graph_rel_model_path2subgraph_rel_model_path(
+                original_graph_rel_model_path,
+                [subgraph_rel_model_path] * len(subgraph_ranges),
+            )
+
+    def _collect_original_graph_rel_model_path2subgraph_rel_model_path(
+        self,
+        original_graph_rel_model_path: str,
+        subgraph_rel_model_paths: list[str],
+    ):
+        old = self.original_graph_rel_model_path2subgraph_rel_model_paths.get(
+            original_graph_rel_model_path, []
+        )
+        self.original_graph_rel_model_path2subgraph_rel_model_paths[
+            original_graph_rel_model_path
+        ] = [
+            *old,
+            *subgraph_rel_model_paths,
+        ]
+
+    def _collect_original_graph_rel_model_path2ranges(
+        self, original_graph_rel_model_path, subgraph_ranges
+    ):
+        old_ranges = self.original_graph_rel_model_path2ranges.get(
+            original_graph_rel_model_path, []
+        )
+        self.original_graph_rel_model_path2ranges[original_graph_rel_model_path] = [
+            *old_ranges,
+            *subgraph_ranges,
+        ]
+
+    def END(self, rel_model_paths: list[str]):
+        for (
+            original_graph_rel_model_path,
+            subgraph_ranges,
+        ) in self.original_graph_rel_model_path2ranges.items():
+            subgraph_rel_model_paths = (
+                self.original_graph_rel_model_path2subgraph_rel_model_paths[
+                    original_graph_rel_model_path
+                ]
+            )
+            self._save_json(
+                original_graph_rel_model_path, subgraph_ranges, subgraph_rel_model_paths
+            )
+
+    def _save_json(
+        self, original_graph_rel_model_path, subgraph_ranges, subgraph_rel_model_paths
+    ):
+        model_dir = Path(self.config["output_dir"]) / original_graph_rel_model_path
+        model_dir.mkdir(parents=True, exist_ok=True)
+        ranges_json = self._get_ranges_json(subgraph_ranges)
+        paths_json = self._get_paths_json(subgraph_rel_model_paths)
+        json_obj = {**ranges_json, **paths_json}
+        json_str = json.dumps(json_obj, indent=4)
+        (model_dir / self.config["output_json_file_name"]).write_text(json_str)
+
+    def _get_paths_json(self, subgraph_rel_model_paths: list[str]):
+        json_obj = {
+            self.config[
+                "output_json_subgraph_rel_model_path_key"
+            ]: subgraph_rel_model_paths
+        }
+        return json_obj
+
+    def _get_ranges_json(self, subgraph_ranges: list[(int, int)]):
+        json_obj = {self.config["output_json_key"]: subgraph_ranges}
+        return json_obj

graph_net/sample_pass/sample_pass.py

@@ -61,7 +61,7 @@ def _make_config_by_config_declare(self, config):
             class_name = type(self).__name__
             assert name in mut_config, f"{name=} {class_name=}"
 
-        def get_extra_config_fields():
+        def get_undefined_config_fields():
             return set(name for name, _ in mut_config.items()) - set(
                 name for name, _ in sig.parameters.items()
             )
@@ -71,10 +71,10 @@ def get_extra_config_fields():
             for _, param in sig.parameters.items()
         )
         if no_varadic_keyword:
-            no_extra_config_fields = all(
+            no_undefined_config_fields = all(
                 name in sig.parameters for name, _ in mut_config.items()
             )
-            assert no_extra_config_fields, f"{get_extra_config_fields()=}"
+            assert no_undefined_config_fields, f"{get_undefined_config_fields()=}"
         return mut_config
 
     def _complete_default(self, name, param, mut_config):

graph_net/test/group_ranges_from_subgraph_sources_test.sh

@@ -0,0 +1,15 @@
+#!/bin/bash
+
+GRAPH_NET_ROOT=$(python3 -c "import graph_net; import os; print(os.path.dirname(os.path.dirname(graph_net.__file__)))")
+
+python3 -m graph_net.apply_sample_pass \
+    --model-path-list "$GRAPH_NET_ROOT/graph_net/test/workspace_group_ranges_from_subgraph_sources/sample_list.txt" \
+    --sample-pass-file-path "$GRAPH_NET_ROOT/graph_net/sample_pass/group_ranges_from_subgraph_sources.py" \
+    --sample-pass-class-name GroupRangesFromSubgraphSources \
+    --sample-pass-config $(base64 -w 0 <<EOF
+{
+    "subgraph_model_path_prefix": "$GRAPH_NET_ROOT/graph_net/test/workspace_group_ranges_from_subgraph_sources",
+    "output_dir": "/tmp/workspace_group_ranges_from_subgraph_sources"
+}
+EOF
+)

graph_net/test/shape_propagator_test.sh

@@ -0,0 +1,16 @@
+#!/bin/bash
+
+GRAPH_NET_ROOT=$(python3 -c "import graph_net; import os; print(os.path.dirname(os.path.dirname(graph_net.__file__)))")
+
+python3 -m graph_net.apply_sample_pass \
+    --model-path-list "$GRAPH_NET_ROOT/graph_net/test/workspace_subgraph_input_producer_indexes_generator/sample_list.txt" \
+    --sample-pass-file-path "$GRAPH_NET_ROOT/graph_net/torch/sample_pass/shape_propagator.py" \
+    --sample-pass-class-name ShapePropagator \
+    --sample-pass-config $(base64 -w 0 <<EOF
+{
+    "resume": false,
+    "model_path_prefix": "$GRAPH_NET_ROOT",
+    "output_dir": "/tmp/workspace_shape_propagator"
+}
+EOF
+)

graph_net/test/subgraph_input_producer_indexes_generator_test.sh

@@ -0,0 +1,19 @@
+#!/bin/bash
+
+GRAPH_NET_ROOT=$(python3 -c "import graph_net; import os; print(os.path.dirname(os.path.dirname(graph_net.__file__)))")
+
+python3 -m graph_net.apply_sample_pass \
+    --model-path-list "$GRAPH_NET_ROOT/graph_net/test/workspace_subgraph_input_producer_indexes_generator/sample_list.txt" \
+    --sample-pass-file-path "$GRAPH_NET_ROOT/graph_net/torch/sample_pass/subgraph_input_producer_indexes_generator.py" \
+    --sample-pass-class-name SubgraphInputProducerIndexesGenerator \
+    --sample-pass-config $(base64 -w 0 <<EOF
+{
+    "resume": false,
+    "model_path_prefix": "$GRAPH_NET_ROOT",
+    "subgraph_ranges_json_root": "$GRAPH_NET_ROOT/graph_net/test/workspace_subgraph_input_producer_indexes_generator",
+    "subgraph_ranges_json_file_name": "grouped_ranges_from_subgraph_sources.json",
+    "subgraph_ranges_json_key": "grouped_ranges_from_subgraph_sources",
+    "output_dir": "/tmp/workspace_subgraph_input_producer_indexes_generator"
+}
+EOF
+)

graph_net/test/workspace_group_ranges_from_subgraph_sources/sample_list.txt

@@ -0,0 +1,5 @@
+samples/timm/mobilenetv4_conv_aa_large.e230_r384_in12k/_decomposed/mobilenetv4_conv_aa_large.e230_r384_in12k_start33_end38_6
+samples/timm/mobilenetv4_conv_aa_large.e230_r384_in12k/_decomposed/mobilenetv4_conv_aa_large.e230_r384_in12k_start55_end60_10
+samples/timm/mobilenetv4_conv_aa_large.e230_r384_in12k/_decomposed/mobilenetv4_conv_aa_large.e230_r384_in12k_start60_end64_11
+samples/timm/convnextv2_base.fcmae_ft_in1k/_decomposed/convnextv2_base.fcmae_ft_in1k_start287_end297_32
+samples/timm/convnextv2_base.fcmae_ft_in1k/_decomposed/convnextv2_base.fcmae_ft_in1k_start401_end411_44

graph_net/test/workspace_group_ranges_from_subgraph_sources/samples/timm/convnextv2_base.fcmae_ft_in1k/_decomposed/convnextv2_base.fcmae_ft_in1k_start287_end297_32/subgraph_sources.json

@@ -0,0 +1,8 @@
+{
+    "samples/timm/convnextv2_base.fcmae_ft_in1k": [
+        [
+            287,
+            297
+        ]
+    ]
+}

graph_net/test/workspace_group_ranges_from_subgraph_sources/samples/timm/convnextv2_base.fcmae_ft_in1k/_decomposed/convnextv2_base.fcmae_ft_in1k_start401_end411_44/subgraph_sources.json

@@ -0,0 +1,8 @@
+{
+    "samples/timm/convnextv2_base.fcmae_ft_in1k": [
+        [
+            401,
+            411
+        ]
+    ]
+}

graph_net/test/workspace_group_ranges_from_subgraph_sources/samples/timm/mobilenetv4_conv_aa_large.e230_r384_in12k/_decomposed/mobilenetv4_conv_aa_large.e230_r384_in12k_start33_end38_6/subgraph_sources.json

@@ -0,0 +1,8 @@
+{
+    "samples/timm/mobilenetv4_conv_aa_large.e230_r384_in12k": [
+        [
+            33,
+            38
+        ]
+    ]
+}

graph_net/test/workspace_group_ranges_from_subgraph_sources/samples/timm/mobilenetv4_conv_aa_large.e230_r384_in12k/_decomposed/mobilenetv4_conv_aa_large.e230_r384_in12k_start55_end60_10/subgraph_sources.json

@@ -0,0 +1,8 @@
+{
+    "samples/timm/mobilenetv4_conv_aa_large.e230_r384_in12k": [
+        [
+            55,
+            60
+        ]
+    ]
+}

graph_net/test/workspace_group_ranges_from_subgraph_sources/samples/timm/mobilenetv4_conv_aa_large.e230_r384_in12k/_decomposed/mobilenetv4_conv_aa_large.e230_r384_in12k_start60_end64_11/subgraph_sources.json

@@ -0,0 +1,8 @@
+{
+    "samples/timm/mobilenetv4_conv_aa_large.e230_r384_in12k": [
+        [
+            60,
+            64
+        ]
+    ]
+}

graph_net/test/workspace_subgraph_input_producer_indexes_generator/sample_list.txt

@@ -0,0 +1,2 @@
+samples/timm/mobilenetv4_conv_aa_large.e230_r384_in12k
+samples/timm/convnextv2_base.fcmae_ft_in1k

graph_net/test/workspace_subgraph_input_producer_indexes_generator/samples/timm/convnextv2_base.fcmae_ft_in1k/grouped_ranges_from_subgraph_sources.json

@@ -0,0 +1,16 @@
+{
+    "grouped_ranges_from_subgraph_sources": [
+        [
+            287,
+            297
+        ],
+        [
+            401,
+            411
+        ]
+    ],
+    "subgraph_rel_model_paths": [
+        "samples/timm/convnextv2_base.fcmae_ft_in1k/_decomposed/convnextv2_base.fcmae_ft_in1k_start287_end297_32",
+        "samples/timm/convnextv2_base.fcmae_ft_in1k/_decomposed/convnextv2_base.fcmae_ft_in1k_start401_end411_44"
+    ]
+}

graph_net/test/workspace_subgraph_input_producer_indexes_generator/samples/timm/mobilenetv4_conv_aa_large.e230_r384_in12k/grouped_ranges_from_subgraph_sources.json

@@ -0,0 +1,21 @@
+{
+    "grouped_ranges_from_subgraph_sources": [
+        [
+            33,
+            38
+        ],
+        [
+            55,
+            60
+        ],
+        [
+            60,
+            64
+        ]
+    ],
+    "subgraph_rel_model_paths": [
+        "samples/timm/mobilenetv4_conv_aa_large.e230_r384_in12k/_decomposed/mobilenetv4_conv_aa_large.e230_r384_in12k_start33_end38_6",
+        "samples/timm/mobilenetv4_conv_aa_large.e230_r384_in12k/_decomposed/mobilenetv4_conv_aa_large.e230_r384_in12k_start55_end60_10",
+        "samples/timm/mobilenetv4_conv_aa_large.e230_r384_in12k/_decomposed/mobilenetv4_conv_aa_large.e230_r384_in12k_start60_end64_11"
+    ]
+}