Summary:
This PR propagates where we use first-class modules objects into the compiler. This creates a transitionary state where:
* compiler.cpp creates Graphs where `self` is a Module class and attributes/parameters/buffers/submodules are looked up with `prim::GetAttr`
* GraphExecutor still runs "lowered graphs" where the self object has been removed by a compiler pass `lower_first_class_method`.
* Tracing still creates "lowered graphs", and a pass "lift_lowered_method" creates a first-class method graph for things.
* This PR separates out Method and Function. A script::Function is a pure Graph with no `self` bound. Similar to Python, a script::Method is just a bound `self` and its underlying `script::Function`.
* This PR also separates CompilationUnit from Module. A CompilationUnit is just a list of named script::Functions. Class's have a CompilationUnit holding the class methods, and Modules also have a CompilationUnit holding their Methods. This avoids the weird circular case Module --has a-> Class -> has a -> Module ...
Details:
* In this transitionary state, we maintain two copies of a Graph, first-class module and lowered. Th first-class one has a self argument that is the module's class type. The lowered one is the lowered graph that uses the initial_ivalues inputs.
* When defining lowered methods using `_defined_lowered` we immediately create the first-class equivalent. The reverse is done lazily, creating lowered_methods on demand from the class.
* The two way conversions will be deleted in a future PR when the executor itself runs first-class objects. However this requires more changes to (1) the traces, (2) the python bindings, and (3) the onnx export pass and would make this PR way to large.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/19167
Differential Revision: D14891966
Pulled By: zdevito
fbshipit-source-id: 0b5f03118aa65448a15c7a7818e64089ec93d7ea
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/18469
ghimport-source-id: 73cb8b58f43f10b1dcfca805fd5b25c4fa977632
Stack from [ghstack](https://github.com/ezyang/ghstack):
* **#18469 Create Object that represents a Module**
* #18468 slots with explicit value/setValue make more sense in future patches
* #18467 Make Object hold its ClassType
* #18379 Enforce single parent for script submodules
* #18378 Unify namespace of script::Module
* #18314 Add ability to specialize class types to ArgumentSpec
* #18226 Add Slot type to abstract the raw pointers being used for slots.
This changes the underlying storage for script::Module to hold
a ivalue::Object which has slots for all the parameters and attributes.
NamedIValue and Slot are now merged together into one class Slot that stores
the tuple (ivalue::Object, offset) and can be used to read the name, type,
or value of the slot and also to set the value. This cleans up a bunch
of client uses.
This PR does not actually use the module object in any generated code.
A future PR will switch how code is generated to treat modules as
first class.
Differential Revision: D14613508
fbshipit-source-id: d853a7559f58d244de2ef54a781427fcd1060ed0
Summary:
This refactor lets us track the types of initial values added onto a `Method`. The main motivation for this is the change in `module.cpp`.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/18252
Differential Revision: D14673459
Pulled By: driazati
fbshipit-source-id: 21200180c47f25bb70898771adfb569856e6c34a
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/18379
ghimport-source-id: 9895ecc1ff7897e98853dc00675341f36726e7c7
Stack from [ghstack](https://github.com/ezyang/ghstack):
* **#18379 Enforce single parent for script submodules**
* #18378 Unify namespace of script::Module
* #18314 Add ability to specialize class types to ArgumentSpec
* #18226 Add Slot type to abstract the raw pointers being used for slots.
The assumption that a ScriptModule has a single parent is present in
our serialization format, and likely a few other places. It is not
enforced on creation of script module hierarchies though, meaning that
problems associated with (e.g. replicating a module twice in the output
format) will not be caught until much later in the development cycle.
This patch enforces the property when a submodule is registered.
It also removes NamedModule since it is no longer necessary in this regime.
This will also allow the easy discover of a modules fully-qualified name
without needing to traverse the Module hierarchy.
Differential Revision: D14603722
fbshipit-source-id: 63ab5d0cccf7d66c7833e0adf9023024ca9607cb
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/18226
ghimport-source-id: b9ec8651212875b30971cc6859d2ddec6559ae3a
If modules become first-class IValues, then the slots will no longer be raw pointers but (IValue, index) pairs. This commit inserts the Slot abstraction so that this change can be made in later patches.
Stack from [ghstack](https://github.com/ezyang/ghstack):
* **#18226 Add Slot type to abstract the raw pointers being used for slots.**
Differential Revision: D14542022
fbshipit-source-id: b81d7f4334c983d663e7551bda82df43680d7c5f
Summary:
Allows serialization/loading of attributes (`IValue`s of any type).
* metadata (attribute name, type) is stored in the `model.json`
* The binary format is a subset of the `pickle` module that supports the operations necessary for `IValue`s
* Attributes are serialized in the order they are defined on a module to a list in a single `attributes` file, with submodule attributes coming first. This order directly matches the order attributes are listed in `model.json`
* This can be inspected in Python with `pickle.load()` or with `pickletools` (PyTorch need not be installed for this to work)
* A class is used to store a tensor's index into the tensor table of the model, so to unpickle the file you have to use a custom Unpickler:
```python
class TensorID(object):
def __setstate__(self, id):
self.id = id
class JitUnpickler(pickle.Unpickler):
def find_class(self, module, name):
if module == '__main__' and name == 'TensorID':
return TensorID
JitUnpickler(open("my_model/attributes.pkl", "rb")).load()
```
* pickle format: https://svn.python.org/projects/python/trunk/Lib/pickletools.py
* It currently does not support/guarantee that anything saved out with `pickle` (i.e. if you edit `attributes` with `pickle` directly) instead of our tools will be imported correctly
Also will fix#17683 and fix#16367
Followup Work:
* document format / choice of pickle: #17951
* create an example
* list specializations
* int size specializations, large binputs
* do a first pass over attributes to output only necessary `BINPUT` ops
* attribute reassignment (e.g `self.my_attribute = new_value`)
* `tensor.save("some_checkpoint.pkl")` support with tensors embedded in Pickle file
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17423
Differential Revision: D14470965
Pulled By: driazati
fbshipit-source-id: 6a21a9939efdbe59b4bc57fd31d6d630bab5297e
Summary:
Stack:
⚫ **#17856 [jit] support serialization of classes** [💛](https://our.intern.facebook.com/intern/diff/D14402599/)
Add support for saving/loading TorchScript modules that depend on user-defned classes.
We track class dependencies the same we track tensor constants, then write them
all out such that we can just compile them in order before compiling the module
hierarchy.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17856
Reviewed By: shannonzhu
Differential Revision: D14461599
Pulled By: suo
fbshipit-source-id: 7115f87e069fd00dc8381d7de9997864fef7ea9f
