Michael Suo d4fa84bf5f Properly serialize types that only appear at function input (#47775) ... Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/47775 When serializing graphs, we check every node for named types referenced, so that we can register them as dependencies. We were skipping this check for the graph inputs themselves. Since types used at input are almost always used somewhere in the graph, we never noticed this gap until a user reported an issue with NamedTuples. Test Plan: Imported from OSS Reviewed By: jamesr66a Differential Revision: D24896289 Pulled By: suo fbshipit-source-id: 4ce76816cb7997a7b65e7cea152ea52ed8f27276		2020-11-11 15:27:00 -08:00
..
export_module.cpp	Add error messages and workaround for RET failure of containers with a torch class type (#46543)	2020-10-26 10:46:07 -07:00
export.cpp	[ONNX] Add dim_param support in export with onnx shape inference (#44920)	2020-09-30 21:56:24 -07:00
export.h	[ONNX] Add dim_param support in export with onnx shape inference (#44920)	2020-09-30 21:56:24 -07:00
import_export_constants.h	Get, save, and load module information for each operator (#42133)	2020-08-14 01:25:27 -07:00
import_export_functions.h
import_export_helpers.cpp
import_export_helpers.h
import_legacy.cpp	Generalize constant_table from tensor only to ivalue (#40718)	2020-07-09 09:09:40 -07:00
import_legacy.h
import_source.cpp	Implement JIT Enum type serialization and deserialization (#43460)	2020-08-24 12:04:31 -07:00
import_source.h	Generalize constant_table from tensor only to ivalue (#40718)	2020-07-09 09:09:40 -07:00
import.cpp	Compile less legacy code when BUILD_CAFFE2 is set to False (take 2) (#44453)	2020-09-11 16:27:47 -07:00
import.h
onnx.cpp	[ONNX] Update ONNX shape inference (#43929)	2020-09-14 15:36:19 -07:00
onnx.h	[ONNX] Utilize ONNX shape inference for ONNX exporter (#40628)	2020-08-30 18:35:46 -07:00
pickle.cpp
pickle.h
pickler.cpp	Add an option to getWriteableTensorData to avoid copy CUDA tensor to CPU (#46524)	2020-10-20 08:54:58 -07:00
pickler.h	Add an option to getWriteableTensorData to avoid copy CUDA tensor to CPU (#46524)	2020-10-20 08:54:58 -07:00
python_print.cpp	Properly serialize types that only appear at function input (#47775)	2020-11-11 15:27:00 -08:00
python_print.h	[jit] Speed up saving in case of many classes (#44589)	2020-09-15 15:10:45 -07:00
source_range_serialization_impl.h
source_range_serialization.cpp
source_range_serialization.h
type_name_uniquer.cpp
type_name_uniquer.h
unpickler.cpp	Define the record_stream method in native_functions.yaml (#44301)	2020-10-13 09:15:22 -07:00
unpickler.h	Extract rpc/tensorpipe_utils.{cpp,h} from rpc/utils.{cpp,h} (#44803)	2020-09-18 13:51:43 -07:00