Update Demo Scripts To Use .ptd (retry)

CMakeLists.txt

@@ -248,14 +248,15 @@ cmake_dependent_option(
   "NOT EXECUTORCH_BUILD_ARM_BAREMETAL" OFF
 )
 
-if(EXECUTORCH_BUILD_EXTENSION_FLAT_TENSOR)
+if(EXECUTORCH_BUILD_EXTENSION_TRAINING)
   set(EXECUTORCH_BUILD_EXTENSION_DATA_LOADER ON)
+  set(EXECUTORCH_BUILD_EXTENSION_FLAT_TENSOR ON)
+  set(EXECUTORCH_BUILD_EXTENSION_MODULE ON)
+  set(EXECUTORCH_BUILD_EXTENSION_TENSOR ON)
 endif()
 
-if(EXECUTORCH_BUILD_EXTENSION_TRAINING)
-  set(EXECUTORCH_BUILD_EXTENSION_TENSOR ON)
+if(EXECUTORCH_BUILD_EXTENSION_FLAT_TENSOR)
   set(EXECUTORCH_BUILD_EXTENSION_DATA_LOADER ON)
-  set(EXECUTORCH_BUILD_EXTENSION_MODULE ON)
 endif()
 
 if(EXECUTORCH_BUILD_EXTENSION_MODULE)

extension/training/CMakeLists.txt

@@ -26,7 +26,7 @@ target_include_directories(
 target_include_directories(extension_training PUBLIC ${EXECUTORCH_ROOT}/..)
 target_compile_options(extension_training PUBLIC ${_common_compile_options})
 target_link_libraries(extension_training executorch_core
-    extension_data_loader extension_module extension_tensor)
+    extension_data_loader extension_module extension_tensor extension_flat_tensor)
 
 
 list(TRANSFORM _train_xor__srcs PREPEND "${EXECUTORCH_ROOT}/")

extension/training/examples/XOR/export_model.py

@@ -11,14 +11,14 @@
 import os
 
 import torch
-from executorch.exir import to_edge
+from executorch.exir import ExecutorchBackendConfig, to_edge
 
 from executorch.extension.training.examples.XOR.model import Net, TrainingNet
 from torch.export import export
 from torch.export.experimental import _export_forward_backward
 
 
-def _export_model():
+def _export_model(external_mutable_weights: bool = False):
     net = TrainingNet(Net())
     x = torch.randn(1, 2)
 
@@ -30,7 +30,11 @@ def _export_model():
     # Lower the graph to edge dialect.
     ep = to_edge(ep)
     # Lower the graph to executorch.
-    ep = ep.to_executorch()
+    ep = ep.to_executorch(
+        config=ExecutorchBackendConfig(
+            external_mutable_weights=external_mutable_weights
+        )
+    )
     return ep
 
 
@@ -44,19 +48,27 @@ def main() -> None:
         "--outdir",
         type=str,
         required=True,
-        help="Path to the directory to write xor.pte files to",
+        help="Path to the directory to write xor.pte and xor.ptd files to",
+    )
+    parser.add_argument(
+        "--external",
+        action="store_true",
+        help="Export the model with external weights",
     )
     args = parser.parse_args()
 
-    ep = _export_model()
+    ep = _export_model(args.external)
 
     # Write out the .pte file.
     os.makedirs(args.outdir, exist_ok=True)
     outfile = os.path.join(args.outdir, "xor.pte")
     with open(outfile, "wb") as fp:
-        fp.write(
-            ep.buffer,
-        )
+        ep.write_to_file(fp)
+
+    if args.external:
+        # current infra doesnt easily allow renaming this file, so just hackily do it here.
+        ep._tensor_data["xor"] = ep._tensor_data.pop("_default_external_constant")
+        ep.write_tensor_data_to_file(args.outdir)
 
 
 if __name__ == "__main__":

extension/training/examples/XOR/train.cpp

@@ -23,12 +23,18 @@ using executorch::extension::training::optimizer::SGDOptions;
 using executorch::runtime::Error;
 using executorch::runtime::Result;
 DEFINE_string(model_path, "xor.pte", "Model serialized in flatbuffer format.");
+DEFINE_string(ptd_path, "", "Model weights serialized in flatbuffer format.");
 
 int main(int argc, char** argv) {
   gflags::ParseCommandLineFlags(&argc, &argv, true);
-  if (argc != 1) {
+  if (argc == 0) {
+    ET_LOG(Error, "Please provide a model path.");
+    return 1;
+  } else if (argc > 2) {
     std::string msg = "Extra commandline args: ";
-    for (int i = 1 /* skip argv[0] (program name) */; i < argc; i++) {
+    for (int i = 2 /* skip argv[0] (pte path) and argv[1] (ptd path) */;
+         i < argc;
+         i++) {
       msg += argv[i];
     }
     ET_LOG(Error, "%s", msg.c_str());
@@ -46,7 +52,21 @@ int main(int argc, char** argv) {
   auto loader = std::make_unique<executorch::extension::FileDataLoader>(
       std::move(loader_res.get()));
 
-  auto mod = executorch::extension::training::TrainingModule(std::move(loader));
+  std::unique_ptr<executorch::extension::FileDataLoader> ptd_loader = nullptr;
+  if (!FLAGS_ptd_path.empty()) {
+    executorch::runtime::Result<executorch::extension::FileDataLoader>
+        ptd_loader_res =
+            executorch::extension::FileDataLoader::from(FLAGS_ptd_path.c_str());
+    if (ptd_loader_res.error() != Error::Ok) {
+      ET_LOG(Error, "Failed to open ptd file: %s", FLAGS_ptd_path.c_str());
+      return 1;
+    }
+    ptd_loader = std::make_unique<executorch::extension::FileDataLoader>(
+        std::move(ptd_loader_res.get()));
+  }
+
+  auto mod = executorch::extension::training::TrainingModule(
+      std::move(loader), nullptr, nullptr, nullptr, std::move(ptd_loader));
 
   // Create full data set of input and labels.
   std::vector<std::pair<
@@ -70,7 +90,10 @@ int main(int argc, char** argv) {
   // Get the params and names
   auto param_res = mod.named_parameters("forward");
   if (param_res.error() != Error::Ok) {
-    ET_LOG(Error, "Failed to get named parameters");
+    ET_LOG(
+        Error,
+        "Failed to get named parameters, error: %d",
+        static_cast<int>(param_res.error()));
     return 1;
   }
 
@@ -112,5 +135,6 @@ int main(int argc, char** argv) {
         std::string(param.first.data()), param.second});
   }
 
-  executorch::extension::flat_tensor::save_ptd("xor.ptd", param_map, 16);
+  executorch::extension::flat_tensor::save_ptd(
+      "trained_xor.ptd", param_map, 16);
 }