Update base for Update on "Serializing from_node info in et serializer"

We need to use the from_node informaton in deserialzied exported graph for operator tracing in et.inspector.  this diff updates the serizalier to support serde from_node info.

Differential Revision: [D78293986](https://our.internmc.facebook.com/intern/diff/D78293986/)

[ghstack-poisoned]

Author: Gasoonjia

.ci/docker/ci_commit_pins/pytorch.txt

@@ -1 +1 @@
-ab43fe4bdf5ccd82897f0e982c451a0127bd175e
+2dccff7dcf56b0d168ebfd7ca08bdeca37273c56

examples/demo-apps/apple_ios/LLaMA/LLaMA.xcodeproj/project.pbxproj

@@ -948,7 +948,7 @@
 			isa = XCRemoteSwiftPackageReference;
 			repositoryURL = "https://github.com/pytorch/executorch";
 			requirement = {
-				branch = "swiftpm-0.8.0.20250714";
+				branch = "swiftpm-0.8.0.20250724";
 				kind = branch;
 			};
 		};

examples/models/llama/README.md

@@ -168,7 +168,7 @@ LLAMA_CHECKPOINT=path/to/consolidated.00.pth
 LLAMA_PARAMS=path/to/params.json
 
 python -m extension.llm.export.export_llm \
-  --config examples/models/llamaconfig/llama_bf16.yaml
+  --config examples/models/llamaconfig/llama_bf16.yaml \
   +base.model_class="llama3_2" \
   +base.checkpoint="${LLAMA_CHECKPOINT:?}" \
   +base.params="${LLAMA_PARAMS:?}" \
@@ -186,7 +186,7 @@ LLAMA_QUANTIZED_CHECKPOINT=path/to/spinquant/consolidated.00.pth.pth
 LLAMA_PARAMS=path/to/spinquant/params.json
 
 python -m extension.llm.export.export_llm \
-  --config examples/models/llama/config/llama_xnnpack_spinquant.yaml
+  --config examples/models/llama/config/llama_xnnpack_spinquant.yaml \
   +base.model_class="llama3_2" \
   +base.checkpoint="${LLAMA_QUANTIZED_CHECKPOINT:?}" \
   +base.params="${LLAMA_PARAMS:?}"
@@ -203,7 +203,7 @@ LLAMA_QUANTIZED_CHECKPOINT=path/to/qlora/consolidated.00.pth.pth
 LLAMA_PARAMS=path/to/qlora/params.json
 
 python -m extension.llm.export.export_llm \
-    --config examples/models/llama/config/llama_xnnpack_qat.yaml
+    --config examples/models/llama/config/llama_xnnpack_qat.yaml \
     +base.model_class="llama3_2" \
     +base.checkpoint="${LLAMA_QUANTIZED_CHECKPOINT:?}" \
     +base.params="${LLAMA_PARAMS:?}" \
@@ -219,15 +219,16 @@ You can export and run the original Llama 3 8B instruct model.
 2. Export model and generate `.pte` file
 ```
 python -m extension.llm.export.export_llm \
-    --config examples/models/llama/config/llama_q8da4w.yaml
-    +base.model_clas="llama3"
+    --config examples/models/llama/config/llama_q8da4w.yaml \
+    +base.model_class="llama3" \
     +base.checkpoint=<consolidated.00.pth.pth> \
     +base.params=<params.json>
 ```
-    Due to the larger vocabulary size of Llama 3, we recommend quantizing the embeddings with `quantization.embedding_quantize=\'4,32\'` as shown above to further reduce the model size.
 
+Due to the larger vocabulary size of Llama 3, we recommend quantizing the embeddings with `quantization.embedding_quantize=\'4,32\'` as shown above to further reduce the model size.
 
-    If you're interested in deploying on non-CPU backends, [please refer the non-cpu-backend section](non_cpu_backends.md)
+
+If you're interested in deploying on non-CPU backends, [please refer the non-cpu-backend section](non_cpu_backends.md)
 
 ## Step 3: Run on your computer to validate
 
@@ -450,7 +451,7 @@ python -m examples.models.llama.eval_llama \
 	-d <checkpoint dtype> \
 	--tasks mmlu \
 	--num_fewshot 5 \
-	--max_seq_len <max sequence length>
+	--max_seq_len <max sequence length> \
 	--max_context_len <max context length>
 ```
 

examples/models/llava/runner/llava_runner.h

@@ -63,7 +63,7 @@ class ET_EXPERIMENTAL LlavaRunner
       bool echo = true) override;
 
  private:
-  inline static const std::string kPresetPrompt =
+  inline static const char* kPresetPrompt =
       "A chat between a curious human and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the human's questions. USER: ";
 };
 

examples/openvino/aot_optimize_and_infer.py

@@ -278,7 +278,7 @@ def transform_fn(x):
             return x[0]
 
         quantized_model = quantize_model(
-            cast(torch.fx.GraphModule, aten_dialect.module()),
+            cast(torch.fx.GraphModule, aten_dialect.module()),  # type: ignore[redundant-cast]
             calibration_dataset,
             subset_size=subset_size,
             transform_fn=transform_fn,

extension/benchmark/apple/Benchmark/Benchmark.xcodeproj/project.pbxproj

@@ -7,6 +7,7 @@
 	objects = {
 
 /* Begin PBXBuildFile section */
+		0314AE3A2E2AAEE700DDE821 /* executorch_llm in Frameworks */ = {isa = PBXBuildFile; productRef = 0314AE392E2AAEE700DDE821 /* executorch_llm */; };
 		032A73CA2CAFBA8600932D36 /* LLaMATests.mm in Sources */ = {isa = PBXBuildFile; fileRef = 032A73C82CAFBA8600932D36 /* LLaMATests.mm */; };
 		0351D9D72CAFC9A200607121 /* Resources in Resources */ = {isa = PBXBuildFile; fileRef = 03C7FA322C8AA24200E6E9AE /* Resources */; };
 		03B0118E2CAC567900054791 /* DynamicTestCase.m in Sources */ = {isa = PBXBuildFile; fileRef = 03B0118C2CAC567900054791 /* DynamicTestCase.m */; };
@@ -18,7 +19,6 @@
 		03F181502D7262FC0058BDF9 /* backend_mps in Frameworks */ = {isa = PBXBuildFile; productRef = 03F1814F2D7262FC0058BDF9 /* backend_mps */; };
 		03F181522D7262FC0058BDF9 /* backend_xnnpack in Frameworks */ = {isa = PBXBuildFile; productRef = 03F181512D7262FC0058BDF9 /* backend_xnnpack */; };
 		03F181542D7262FC0058BDF9 /* executorch in Frameworks */ = {isa = PBXBuildFile; productRef = 03F181532D7262FC0058BDF9 /* executorch */; };
-		0314AE3A2E2AAEE700DDE821 /* executorch_llm in Frameworks */ = {isa = PBXBuildFile; productRef = 0314AE392E2AAEE700DDE821 /* executorch_llm */; };
 		03F181562D7262FC0058BDF9 /* kernels_llm in Frameworks */ = {isa = PBXBuildFile; productRef = 03F181552D7262FC0058BDF9 /* kernels_llm */; };
 		03F181582D7262FC0058BDF9 /* kernels_optimized in Frameworks */ = {isa = PBXBuildFile; productRef = 03F181572D7262FC0058BDF9 /* kernels_optimized */; };
 		03F1815C2D7262FC0058BDF9 /* kernels_quantized in Frameworks */ = {isa = PBXBuildFile; productRef = 03F1815B2D7262FC0058BDF9 /* kernels_quantized */; };
@@ -41,7 +41,6 @@
 		03B0118C2CAC567900054791 /* DynamicTestCase.m */ = {isa = PBXFileReference; lastKnownFileType = sourcecode.c.objc; path = DynamicTestCase.m; sourceTree = "<group>"; };
 		03B0118F2CAD114E00054791 /* ResourceTestCase.h */ = {isa = PBXFileReference; lastKnownFileType = sourcecode.c.h; path = ResourceTestCase.h; sourceTree = "<group>"; };
 		03B011902CAD114E00054791 /* ResourceTestCase.m */ = {isa = PBXFileReference; lastKnownFileType = sourcecode.c.objc; path = ResourceTestCase.m; sourceTree = "<group>"; };
-		03B019502C8A80D30044D558 /* Tests.xcconfig */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = text.xcconfig; path = Tests.xcconfig; sourceTree = "<group>"; };
 		03B2D3642C8A515A0046936E /* Benchmark.app */ = {isa = PBXFileReference; explicitFileType = wrapper.application; includeInIndex = 0; path = Benchmark.app; sourceTree = BUILT_PRODUCTS_DIR; };
 		03B2D3672C8A515A0046936E /* App.swift */ = {isa = PBXFileReference; lastKnownFileType = sourcecode.swift; path = App.swift; sourceTree = "<group>"; };
 		03B2D36D2C8A515B0046936E /* App.entitlements */ = {isa = PBXFileReference; lastKnownFileType = text.plist.entitlements; path = App.entitlements; sourceTree = "<group>"; };
@@ -131,7 +130,6 @@
 				032A73C92CAFBA8600932D36 /* LLaMA */,
 				03E7E6782CBDC1C900205E71 /* CoreMLTests.mm */,
 				03B2D3792C8A515C0046936E /* GenericTests.mm */,
-				03B019502C8A80D30044D558 /* Tests.xcconfig */,
 				037C96A02C8A570B00B3DF38 /* Tests.xctestplan */,
 			);
 			path = Tests;
@@ -438,7 +436,6 @@
 		};
 		03B2D38D2C8A515C0046936E /* Debug */ = {
 			isa = XCBuildConfiguration;
-			baseConfigurationReference = 03B019502C8A80D30044D558 /* Tests.xcconfig */;
 			buildSettings = {
 				ALWAYS_EMBED_SWIFT_STANDARD_LIBRARIES = YES;
 				BUNDLE_LOADER = "$(TEST_HOST)";
@@ -450,6 +447,7 @@
 				MACOSX_DEPLOYMENT_TARGET = 12.0;
 				MARKETING_VERSION = 1.0;
 				OTHER_CODE_SIGN_FLAGS = "--deep";
+				OTHER_LDFLAGS = "-all_load";
 				PRODUCT_BUNDLE_IDENTIFIER = org.pytorch.executorch.BenchmarkTests;
 				PRODUCT_NAME = "$(TARGET_NAME)";
 				REGISTER_APP_GROUPS = NO;
@@ -465,7 +463,6 @@
 		};
 		03B2D38E2C8A515C0046936E /* Release */ = {
 			isa = XCBuildConfiguration;
-			baseConfigurationReference = 03B019502C8A80D30044D558 /* Tests.xcconfig */;
 			buildSettings = {
 				ALWAYS_EMBED_SWIFT_STANDARD_LIBRARIES = YES;
 				BUNDLE_LOADER = "$(TEST_HOST)";
@@ -477,6 +474,7 @@
 				MACOSX_DEPLOYMENT_TARGET = 12.0;
 				MARKETING_VERSION = 1.0;
 				OTHER_CODE_SIGN_FLAGS = "--deep";
+				OTHER_LDFLAGS = "-all_load";
 				PRODUCT_BUNDLE_IDENTIFIER = org.pytorch.executorch.BenchmarkTests;
 				PRODUCT_NAME = "$(TARGET_NAME)";
 				REGISTER_APP_GROUPS = NO;
@@ -530,6 +528,10 @@
 /* End XCLocalSwiftPackageReference section */
 
 /* Begin XCSwiftPackageProductDependency section */
+		0314AE392E2AAEE700DDE821 /* executorch_llm */ = {
+			isa = XCSwiftPackageProductDependency;
+			productName = executorch_llm;
+		};
 		03F1814D2D7262FC0058BDF9 /* backend_coreml */ = {
 			isa = XCSwiftPackageProductDependency;
 			productName = backend_coreml;
@@ -546,10 +548,6 @@
 			isa = XCSwiftPackageProductDependency;
 			productName = executorch;
 		};
-		0314AE392E2AAEE700DDE821 /* executorch_llm */ = {
-			isa = XCSwiftPackageProductDependency;
-			productName = executorch_llm;
-		};
 		03F181552D7262FC0058BDF9 /* kernels_llm */ = {
 			isa = XCSwiftPackageProductDependency;
 			productName = kernels_llm;

extension/benchmark/apple/Benchmark/Tests/Tests.xcconfig

@@ -71,7 +71,7 @@ def python_is_compatible():
 #
 # NOTE: If you're changing, make the corresponding change in .ci/docker/ci_commit_pins/pytorch.txt
 # by picking the hash from the same date in https://hud.pytorch.org/hud/pytorch/pytorch/nightly/
-NIGHTLY_VERSION = "dev20250716"
+NIGHTLY_VERSION = "dev20250723"
 
 
 def install_requirements(use_pytorch_nightly):

install_requirements.py

@@ -1,78 +1 @@
-#ifndef C10_MACROS_EXPORT_H_
-#define C10_MACROS_EXPORT_H_
-
-#ifndef C10_USING_CUSTOM_GENERATED_MACROS
-#include <torch/headeronly/macros/cmake_macros.h>
-#endif // C10_USING_CUSTOM_GENERATED_MACROS
-
 #include <torch/headeronly/macros/Export.h>
-
-// This one is being used by libtorch.so
-#ifdef CAFFE2_BUILD_MAIN_LIB
-#define TORCH_API C10_EXPORT
-#else
-#define TORCH_API C10_IMPORT
-#endif
-
-// You may be wondering: Whose brilliant idea was it to split torch_cuda into
-// two pieces with confusing names?
-// Once upon a time, there _was_ only TORCH_CUDA_API. All was happy until we
-// tried to compile PyTorch for CUDA 11.1, which ran into relocation marker
-// issues when linking big binaries.
-// (https://github.com/pytorch/pytorch/issues/39968) We had two choices:
-//    (1) Stop supporting so many GPU architectures
-//    (2) Do something else
-// We chose #2 and decided to split the behemoth that was torch_cuda into two
-// smaller libraries, one with most of the core kernel functions (torch_cuda_cu)
-// and the other that had..well..everything else (torch_cuda_cpp). The idea was
-// this: instead of linking our static libraries (like the hefty
-// libcudnn_static.a) with another huge library, torch_cuda, and run into pesky
-// relocation marker issues, we could link our static libraries to a smaller
-// part of torch_cuda (torch_cuda_cpp) and avoid the issues.
-
-// libtorch_cuda_cu.so
-#ifdef TORCH_CUDA_CU_BUILD_MAIN_LIB
-#define TORCH_CUDA_CU_API C10_EXPORT
-#elif defined(BUILD_SPLIT_CUDA)
-#define TORCH_CUDA_CU_API C10_IMPORT
-#endif
-
-// libtorch_cuda_cpp.so
-#ifdef TORCH_CUDA_CPP_BUILD_MAIN_LIB
-#define TORCH_CUDA_CPP_API C10_EXPORT
-#elif defined(BUILD_SPLIT_CUDA)
-#define TORCH_CUDA_CPP_API C10_IMPORT
-#endif
-
-// libtorch_cuda.so (where torch_cuda_cu and torch_cuda_cpp are a part of the
-// same api)
-#ifdef TORCH_CUDA_BUILD_MAIN_LIB
-#define TORCH_CUDA_CPP_API C10_EXPORT
-#define TORCH_CUDA_CU_API C10_EXPORT
-#elif !defined(BUILD_SPLIT_CUDA)
-#define TORCH_CUDA_CPP_API C10_IMPORT
-#define TORCH_CUDA_CU_API C10_IMPORT
-#endif
-
-#if defined(TORCH_HIP_BUILD_MAIN_LIB)
-#define TORCH_HIP_CPP_API C10_EXPORT
-#define TORCH_HIP_API C10_EXPORT
-#else
-#define TORCH_HIP_CPP_API C10_IMPORT
-#define TORCH_HIP_API C10_IMPORT
-#endif
-
-#if defined(TORCH_XPU_BUILD_MAIN_LIB)
-#define TORCH_XPU_API C10_EXPORT
-#else
-#define TORCH_XPU_API C10_IMPORT
-#endif
-
-// Enums only need to be exported on windows for non-CUDA files
-#if defined(_WIN32) && defined(__CUDACC__)
-#define C10_API_ENUM C10_API
-#else
-#define C10_API_ENUM
-#endif
-
-#endif // C10_MACROS_EXPORT_H_