Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/70212
Use DynamicType instead of ListType all over the place in Lite Interpreter. Namely we need to modify the following places:
1. Type parser which produces the Type constants.
2. IValue::type() which returns reflected Type from IValues.
3. Helper functions to construct the container value.
4. Typechecks which test whether a type instance is a particular container type.
ghstack-source-id: 146818619
Test Plan: CI
Reviewed By: iseeyuan
Differential Revision: D33176931
fbshipit-source-id: 9144787f5fc4778538e5c665946974eb6171a2e6
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/70202
Use DynamicType instead of DictType all over the place in Lite Interpreter. Namely we need to modify the following places:
1. Type parser which produces the Type constants.
2. IValue::type() which returns reflected Type from IValues.
3. Helper functions to construct the container value.
4. Typechecks which test whether a type instance is a particular container type.
ghstack-source-id: 146735648
Test Plan: no behavior change.
Reviewed By: iseeyuan
Differential Revision: D33137257
fbshipit-source-id: 971bf431658c422ea9353cc32cdab66e98876e9d
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/70338
Today Unpickler is used by both server and mobile for deserializing model, and it always fallback to mobile parser when there's no type resolver provided by user. However this is not intended as server and mobile type parser supports different things. In this diff we provide a default fallback using script parser and opt it out for all mobile cases.
ghstack-source-id: 146727330
(Note: this ignores all push blocking failures!)
Test Plan: CI
Reviewed By: iseeyuan
Differential Revision: D33284352
fbshipit-source-id: 997c4f110b36eee6596e8f23f6a87bf91a4197ed
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/68136
DynamicType is an extension to existing server JIT types. Today using normal server types on Edge is a bit problematic because in embedded environments we don't need the full spectrum of types but we still build with these unneeded dependencies.
Is it possible to just get rid of unneeded JIT types from Edge builds? It's not easy to do so at this moment. For example, on Edge we don't support Union type, but we have to pull in the dependency of Union type because Optional type is being supported which inherits from Union type, so Union type has to be included in the build. Although we could split Union type and Optional type, it could be argued that the root cause is every time we use anything inheriting from `c10::Type`, we don't have the direct evidence of how much dependency we pull in, because we do virtual calls and we don't know what exactly we're calling with server JIT types. If we don't know, it's highly possible that the linker doesn't know either so it cannot effectively strip unused methods.
To address this problem, one option is to implement a separate `DynamicType` which has simpler behavior and doesn't store different types as different symbols in binary but rather raw data (or "tag"). This could increase the binary size by several KBs, so I included several binary size reductions in the same stack, hoping at least we don't regress the binary size.
Currently `DynamicType` inherits from `c10::Type` because I want to reduce the migration cost of `DynamicType` by making it interfacing with existing server JIT types. In the future `DynamicType` should be implemented as a separate class without relying on `c10::Type` to make things both simpler and leaner.
ghstack-source-id: 146670522
Test Plan: in the next diff.
Reviewed By: VitalyFedyunin
Differential Revision: D32264615
fbshipit-source-id: 180eb0998a14eacc1d8b28db39870d84fcc17d5b
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/67787
First noticed in https://fb.workplace.com/groups/pytorch.edge.team/posts/952737705280969/ - basically one of the speech models has ~400 0 byte tensor files, so we're basically paying the cost of looking it up in the archive and reading nothing from it.
Turns out that there's a fairly simple fix to avoid reading a 0 byte tensor. Once we notice that it's 0 bytes, just use the default `DataPtr` instead to initializing it with 0 bytes read in from the input file stream.
ghstack-source-id: 142025211
Test Plan: CI and manually ran a couple production mobile models with bundled inputs. CI Will run all prod. mobile mobiles with bundled inputs.
Reviewed By: swolchok
Differential Revision: D32054983
fbshipit-source-id: 919b0cdbc44bccb8f6cfe0da10ff5474af37fd99
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/66693
Passing a `TypePtr` by value causes an unnececssary refcount
bump. We don't need to take ownership, so `const Type&` is all we
need.
I considered providing a compatibility shim that takes `const
TypePtr&`, but doing so is dangerous because a
copy is required to convert from a more specific pointer like
`NoneTypePtr`.
ghstack-source-id: 140737081
Test Plan: CI
Reviewed By: suo
Differential Revision: D31691869
fbshipit-source-id: f766ce3234a28771c2a9ca4c284eb3f96993a3d0
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/64066
I noticed a bunch of time being spent heap-allocating Tuples
in the unpickler. 1-, 2-, and 3-element Tuples are apparently common
enough that they get their own bytecode instructions, so I decided to
try also giving them their own representation. We store up to 3
IValues inline in `Tuple` rather than doing a second heap allocation
for a `std::vector<IValue>`.
ghstack-source-id: 140695395
Test Plan:
Added automated tests for TupleElements.
Pixel 3 before: https://www.internalfb.com/intern/aibench/details/761596366576284
Pixel 3 after: https://www.internalfb.com/intern/aibench/details/591414145082422
We went from 347 ms to 302 ms.
Reviewed By: dhruvbird
Differential Revision: D30592622
fbshipit-source-id: 93625c54c9dca5f765ef6d5c191944179cb281a8
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/64065
It is only safe to mutate Tuple elements if you are the sole owner
of the tuple. The most efficient way to do this, then, is
`std::move(*std::move(tupleIValue).toTuple()).elements()` (the
innermost move allows `IValue::toTuple()` to avoid a refcount bump and
the outermost move allows the element vector to be moved out of the
tuple), but many callsites write simply
`tupleIValue.toTuple().elements()`, which incurs many extra refcount
bumps.
ghstack-source-id: 139468088
Test Plan: CI
Reviewed By: ezyang
Differential Revision: D30592621
fbshipit-source-id: e8312de866de09b9ea2a62e5128cbf403ee16f09
Summary:
This PR is created to replace https://github.com/pytorch/pytorch/pull/53180 PR stack, which has all the review discussions. Reason for needing a replacement is due to a messy Sandcastle issue.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/64234
Reviewed By: gmagogsfm
Differential Revision: D30656444
Pulled By: ansley
fbshipit-source-id: 77536c8bcc88162e2c72636026ca3c16891d669a
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/63974
Saw some time spent in this for model loading, no reason not to move here.
ghstack-source-id: 136760979
Test Plan: Re-profile model loading on devserver; IValue copy ctor time has gone down
Reviewed By: dhruvbird
Differential Revision: D30548923
fbshipit-source-id: 42000f2e18582762b43353cca10ae094833de3b3
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/62794
This pr updates jit serialization to support pickling Sparse COO tensors.
This pr updates message.cpp to support Sparse COO tensors.
A bug was filed a few years ago https://github.com/pytorch/pytorch/issues/30807.
I tested the fix by adding sparse tensor tests to rpc_test.py and dist_autograd_test.py.
cc pietern mrshenli pritamdamania87 zhaojuanmao satgera rohan-varma gqchen aazzolini osalpekar jiayisuse agolynski SciPioneer H-Huang mrzzd cbalioglu gcramer23 gmagogsfm
Test Plan: Imported from OSS
Reviewed By: soulitzer
Differential Revision: D30608848
Pulled By: gcramer23
fbshipit-source-id: 629ba8e4a3d8365875a709c9b87447c7a71204fb
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/59735
1. Fixes ABA storage identity problem during serialization for `torch.package` by keeping reference of serialized storages through lifetime of `PackageExporter` to prevent reuse of memory address. Achieved by extending logic used in solution to mobile's same issue.
2. Adds determinism to naming scheme of serialized storages in export code paths which utilize `tensor_cdata_naming_scheme`(introduced 2nd mapping in `StorageContext`, now maps `storage cdata ptr` -> `unique id`, `unique id` -> `c10::Storage`)
3. Additionally uses presence of a storage in the `StorageContext` instance as marker for if a storage has been serialized or not, removing the need to scan the `PythonStreamWriter` for presence of the storage's serialization file
Test Plan: Imported from OSS
Reviewed By: suo
Differential Revision: D29075276
Pulled By: Lilyjjo
fbshipit-source-id: 15a5c30b1de99c5bd7079388f2db9b6ece2eca12
Summary:
This is an automatic change generated by the following script:
```
#!/usr/bin/env python3
from subprocess import check_output, check_call
import os
def get_compiled_files_list():
import json
with open("build/compile_commands.json") as f:
data = json.load(f)
files = [os.path.relpath(node['file']) for node in data]
for idx, fname in enumerate(files):
if fname.startswith('build/') and fname.endswith('.DEFAULT.cpp'):
files[idx] = fname[len('build/'):-len('.DEFAULT.cpp')]
return files
def run_clang_tidy(fname):
check_call(["python3", "tools/clang_tidy.py", "-c", "build", "-x", fname,"-s"])
changes = check_output(["git", "ls-files", "-m"])
if len(changes) == 0:
return
check_call(["git", "commit","--all", "-m", f"NOLINT stubs for {fname}"])
def main():
git_files = check_output(["git", "ls-files"]).decode("ascii").split("\n")
compiled_files = get_compiled_files_list()
for idx, fname in enumerate(git_files):
if fname not in compiled_files:
continue
if fname.startswith("caffe2/contrib/aten/"):
continue
print(f"[{idx}/{len(git_files)}] Processing {fname}")
run_clang_tidy(fname)
if __name__ == "__main__":
main()
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/56892
Reviewed By: H-Huang
Differential Revision: D27991944
Pulled By: malfet
fbshipit-source-id: 5415e1eb2c1b34319a4f03024bfaa087007d7179
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/54428
Using c10::ArrayRef as the parameter type makes the API more flexible and allows the caller to leverage small-buffer optimizations (e.g. c10::SmallVector, std::array) for performance critical cases.
Test Plan: No behavioral changes. Run the existing unit and integration tests.
Reviewed By: suo
Differential Revision: D27232222
fbshipit-source-id: 7b13bc6bd02257097ca119077028fbccc68cc925
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/51902
These seem like straightforward improvements. (I don't have measurements; feel free to reject if you're skeptical)
ghstack-source-id: 121278775
Test Plan: CI
Reviewed By: qizzzh
Differential Revision: D26322438
fbshipit-source-id: d393a32cc34bb68bc4f804f4b1cc5a8af27763c9
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/50229
`fastmod -m 'cast(<((at|c10)::)?\w+Type>\(\)\s*)->' 'castRaw${1}->'` Presuming it builds, this is a safe change: the
result of `cast()` wasn't being saved anywhere, so we didn't need
it, so we can use a raw pointer instead of a new `shared_ptr`.
ghstack-source-id: 120769170
Test Plan: CI
Reviewed By: SplitInfinity
Differential Revision: D25837494
fbshipit-source-id: 46319100dc0dfc78f6d2b45148207f83481f2ada
Summary:
Add a new device type 'XPU' ('xpu' for lower case) to PyTorch. Changes are needed for code related to device model and kernel dispatch, e.g. DeviceType, Backend and DispatchKey etc.
https://github.com/pytorch/pytorch/issues/48246
Pull Request resolved: https://github.com/pytorch/pytorch/pull/49786
Reviewed By: mrshenli
Differential Revision: D25893962
Pulled By: ezyang
fbshipit-source-id: 7ff0a316ee34cf0ed6fc7ead08ecdeb7df4b0052
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/48868
Building on the previous diff, we can make `toTensor()` return a
`const Tensor&`, which should make it easier to avoid reference
counting.
ghstack-source-id: 119327372
Test Plan: internal benchmarks.
Reviewed By: bwasti
Differential Revision: D25325379
fbshipit-source-id: ca699632901691bcee432f595f75b0a4416d55dd
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/49146
Add support for Storage arguments to IValue and the JIT typing system, and make ops that were blocked on that c10-full.
ghstack-source-id: 118710665
(Note: this ignores all push blocking failures!)
Test Plan: waitforsandcastle
Reviewed By: ezyang
Differential Revision: D25456799
fbshipit-source-id: da14f125af352de5fcf05a83a69ad5a69d5a3b45
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/46827
TensorPipe RPC agent uses JIT pickler/unpickler to serialize/deserialize
tensors. Instead of saving tensors to a file, the agent can directly
invoke `cudaMemcpy` to copy tensors from the sender to the receiver
before calling into JIT unpickler. As a result, before unpickling,
the agent might already have allocated tensors and need to pass
them to the JIT unpickler. Currently, this is done by providing a
`read_record` lambda to unpickler for CPU tensors, but this is
no longer sufficient for zero-copy CUDA tensors, as the unpickler
always allocate the tensor on CPU.
To address the above problem, this commit introduces a `use_storage_device`
flag to unpickler ctor. When this flag is set, the unpickler will
use the device from the `DataPtr` returned by the `read_record`
lambda to override the pickled device information and therefore
achieves zero-copy.
Test Plan: Imported from OSS
Reviewed By: wanchaol
Differential Revision: D24533218
Pulled By: mrshenli
fbshipit-source-id: 35acd33fcfb11b1c724f855048cfd7b2991f8903
Summary:
The record_stream method was hard coded for CUDA device. Define the record_stream in the native_functions.yaml to enable the dynamic dispatch to different end device.
Fixes https://github.com/pytorch/pytorch/issues/36556
Pull Request resolved: https://github.com/pytorch/pytorch/pull/44301
Reviewed By: glaringlee
Differential Revision: D23763954
Pulled By: ezyang
fbshipit-source-id: e6d24f5e7892b56101fa858a6cad2abc5cdc4293
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/43088
Create quantized module that the user can use to perform embedding bag quantization
The module uses the EmbeddingPackedParams to store the weights which can be serialized /deserialized
using TorchBind custom classes (C++ get/setstate code)
Following PR will add support for `from_float` to convert from float to quantized module
Test Plan:
python test/test_quantization.py TestDynamicQuantizedModule.test_embedding_bag_api
Imported from OSS
Reviewed By: vkuzo
Differential Revision: D23167519
fbshipit-source-id: 029d7bb44debf78c4ef08bfebf267580ed94d033
Summary:
* Add EnumType and AnyEnumType as first-class jit type
* Add Enum-typed IValue
* Enhanced aten::eq to support Enum
Supported:
Enum-typed function targuments
using Enum type and comparing them
TODO:
Add PyThon sugared value for Enum
Support getting name/value attrs of enums
Support Enum-typed return values
Support enum values of different types in same Enum class
Support serialization and deserialization
Pull Request resolved: https://github.com/pytorch/pytorch/pull/41390
Reviewed By: eellison
Differential Revision: D22524388
Pulled By: gmagogsfm
fbshipit-source-id: 1627154a64e752d8457cd53270f3d14aea4b1150
Summary:
Clearly expressing a type is inferred by PyTorch instead of explicitly annotated by user makes many error messages more user-friendly
Currently Type has two string conversion methods. str() for IR printing and python_str() for serialization and error message generation. If we want to include more information in type printing while maintaining serialization/deserialization correctness, we need to split python_str() into annotation_str() and repr_str().
annotation_str is solely responsible for serialization, it strictly matches format of python type annotation. repr_str() is responsible for generating a human-readable error message that includes information like "this type is inferred, not explicitly annotated"
Closes https://github.com/pytorch/pytorch/issues/39449
Pull Request resolved: https://github.com/pytorch/pytorch/pull/39544
Differential Revision: D21978759
Pulled By: gmagogsfm
fbshipit-source-id: 733566f5a62e748b5ca4bb3c5943ebb6d5b664d0
Summary:
**Summary**
This commit adds support for seralization and deserialization of
`ScriptModules` that have been lowered to a specific backend. Nothing
special was required to accomplish this, other than removing some code
in `unpickler.cpp` that guarded against the deserialization of `Any`
type objects. Now that lists and dicts are tagged with their types
during serialization, this check is no longer necessary.
**Test Plan**
This commit adds a unit test for testing that a lowered module still
produces the same results as Python and regular JIT after saving and
loading.
**Fixes**
This pull request fixes part of https://github.com/pytorch/pytorch/issues/37841.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/38893
Differential Revision: D21825813
Pulled By: SplitInfinity
fbshipit-source-id: 77a7b84504e0dddf14c89b3ed5dd6b438c086f66
