Summary:
On the way to #10774
This PR adds advanced indexing with tensors.
The approach is to desugar advanced indexing into an at::index op.
This is exactly how normal pytorch does it.
[(I used this code as reference)](https://github.com/pytorch/pytorch/blob/master/torch/csrc/autograd/python_variable_indexing.cpp)
Supporting sequences is a little tricky because JIT script doesn't have
an easy way to turn arbitrary n-dimensional python lists into a tensor
(it would be easy if we supported `torch.tensor`), so that'll come
in a future PR.
cc jamesr66a zdevito
Pull Request resolved: https://github.com/pytorch/pytorch/pull/10862
Differential Revision: D9659449
Pulled By: zou3519
fbshipit-source-id: 56d293720d44c0fd27909e18327ab3985ddfced6
Summary:
This lets you compile builtin functions from C++ without having a dependence on Python
```cpp
auto module = torch::jit::compile(JIT"(
def my_script_method(x, y):
return torch.relu(x) + y
)");
IValue result = module->run_method("my_script_method", 1, 2);
```
goldsborough zdevito apaszke
Pull Request resolved: https://github.com/pytorch/pytorch/pull/10847
Differential Revision: D9543461
Pulled By: driazati
fbshipit-source-id: 6160dae094030ca144a0df93cb9f26aa78c8cf27
Summary:
Adding short circuit evaluation to AND or OR. The second expression of and AND or OR gets lifted into an if branch, which is conditionally evaluated.
BatchOps was using the expression `dims = dims1 or dims2`, where dims is often an empty tensor. This nows throws an error, because dims1 gets cast to a boolean, and you can't convert an empty tensor to a scalar. It now matches the behavior of pytorch in python.
One thing that came up is if the second expression in an and/or in python gets returned, it does not get coerced to a boolean.
`tensor == (False or tensor)`
`tensor == (True and tensor)`
We do not currently support this.
edit: wording
Pull Request resolved: https://github.com/pytorch/pytorch/pull/11116
Differential Revision: D9618168
Pulled By: eellison
fbshipit-source-id: 93b202be2f222d41f85d38d9c95f04d1749e8343
Summary:
This places all constants in the entry block of the graph, and de-duplicates them.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/10231
Differential Revision: D9601501
Pulled By: resistor
fbshipit-source-id: daa10ed8c99e9894830d6f3e5d65c8d3ab5ea899
Summary:
Operators like aten::chunk used to return a number of tensors, but
now return a list. To make it easier to do shape prop through
aten::chunk and fuse it, I've also introduced prim::ConstantChunk,
which behaves like the previous implementation (has a variable length
output list).
The downside of this PR is that the introduction of more lists to the IR causes the LSTM and MiLSTM graphs to be considered as non-differentiable by the graph executor. I verified that they are still optimize correctly, and my next patch (that changes how the specializations/differentiation works) will restore those.
zdevito
Pull Request resolved: https://github.com/pytorch/pytorch/pull/10949
Reviewed By: zdevito
Differential Revision: D9556823
Pulled By: apaszke
fbshipit-source-id: 33e63b17fc7247cac6cfc05eb7eb9bf069b499ee
Summary:
TODO: integrate into torch.onnx.export -- separate PR
*Problem:* We have a facility to trace PyTorch operations on Python code, but there are several failure modes where the trace is not representative of the actual underlying computation:
* The tracer encountered dynamic control flow
* Some computation escaped the tracer, and appeared as a Constant tensor node in the graph
* Some stateful function was traced, e.g. someone did an optimization in Python by memoizing function outputs
*Objective*: In an ideal world, this whole process would be automated and the user can trust that the system will magically capture the intended semantics from the program. Realistically speaking, we will likely have to settle with a human-in-the-loop error reporting system, allowing for the user to identify problems and modify the source code to allow for tracing.
*Stage 1* (this PR): Output-level checking & graph diff. torch.jit.trace gains a kwarg 'check_inputs', which is a list of tuples of input arguments. We will iterate through the list and trace the function again for each set of check inputs. We'll also interpret the original trace with these inputs and compare output values and graphs, printing a diff of the graph if there is a difference.
Examples:
```
torch.jit.trace(torch.rand(3, 4), check_inputs=[(torch.rand(4, 5),)])
def foo(x):
y = torch.arange(0, x.shape[0]).float()
return x + y.unsqueeze(1)
```
```
torch.jit.TracingCheckError: Tracing failed sanity checks!
ERROR: Graphs differed across invocations!
Graph diff:
graph(%0 : Dynamic) {
- %1 : Dynamic = prim::Constant[value= 0 1 2 [ CPULongType{3} ]]()
? ^
+ %1 : Dynamic = prim::Constant[value= 0 1 2 3 [ CPULongType{4} ]]()
? +++ ^
%2 : int = prim::Constant[value=0]()
%3 : Dynamic = aten::_cast_Float(%1, %2)
%4 : int = prim::Constant[value=1]()
%5 : Dynamic = aten::unsqueeze(%3, %4)
%6 : int = prim::Constant[value=1]()
%7 : Dynamic = aten::add(%0, %5, %6)
return (%7);
}
Node diff:
- %1 : Dynamic = prim::Constant[value= 0 1 2 [ CPULongType{3} ]]()
? ^
+ %1 : Dynamic = prim::Constant[value= 0 1 2 3 [ CPULongType{4} ]]()
? +++ ^
Trace source location:
dank.py(5): foo
/Users/jamesreed/onnx-fairseq/pytorch/torch/jit/__init__.py(402): wrapper
dank.py(3): <module>
Check source location:
dank.py(5): foo
/Users/jamesreed/onnx-fairseq/pytorch/torch/jit/__init__.py(281): check_trace
/Users/jamesreed/onnx-fairseq/pytorch/torch/jit/__init__.py(408): wrapper
dank.py(3): <module>
ERROR: Tensor-valued Constant nodes differed in value across invocations. This often indicates that the tracer has encountered untraceable code.
Node:
%1 : Dynamic = prim::Constant[value= 0 1 2 [ CPULongType{3} ]]()
Source Location:
dank.py(5): foo
/Users/jamesreed/onnx-fairseq/pytorch/torch/jit/__init__.py(402): wrapper
dank.py(3): <module>
Comparison exception:
Not equal to tolerance rtol=1e-07, atol=0
(shapes (3,), (4,) mismatch)
x: array([0, 1, 2])
y: array([0, 1, 2, 3])
```
==
```
torch.jit.trace(torch.rand(3, 4), check_inputs=[(torch.rand(3, 4),)])
def foo(x):
y = x.data
return x + y
```
```
torch.jit.TracingCheckError: Tracing failed sanity checks!
ERROR: Traced function outputs do not match the Python function outputs.
ERROR: Tensor-valued Constant nodes differed in value across invocations. This often indicates that the tracer has encountered untraceable code.
Node:
%1 : Dynamic = prim::Constant[value=<Tensor>]()
Source Location:
dank.py(6): foo
/Users/jamesreed/onnx-fairseq/pytorch/torch/jit/__init__.py(402): wrapper
dank.py(3): <module>
Comparison exception:
Not equal to tolerance rtol=1e-07, atol=0
(mismatch 100.0%)
x: array([0.397137, 0.956105, 0.169478, 0.560292, 0.392568, 0.108441,
0.97645 , 0.34412 , 0.951246, 0.793061, 0.557595, 0.770245],
dtype=float32)
y: array([0.243178, 0.315964, 0.972041, 0.0215 , 0.927751, 0.457512,
0.951092, 0.97883 , 0.048688, 0.118066, 0.779345, 0.271272],
dtype=float32)
```
==
```
import torch
torch.jit.trace(torch.rand(3, 4), check_inputs=[(torch.rand(4, 4),)])
def foo(x):
for _ in range(x.size(0)):
x = torch.neg(x)
return x
```
```
torch.jit.TracingCheckError: Tracing failed sanity checks!
ERROR: Traced function outputs do not match the Python function outputs.
ERROR: Graphs differed across invocations!
Graph diff:
graph(%0 : Dynamic) {
%1 : Dynamic = aten::neg(%0)
%2 : Dynamic = aten::neg(%1)
%3 : Dynamic = aten::neg(%2)
+ %4 : Dynamic = aten::neg(%3)
- return (%3);
? ^
+ return (%4);
? ^
}
```
==
```
import torch
def foo(x):
if not hasattr(foo, 'cache'):
foo.cache = torch.neg(x)
return x + foo.cache
traced = torch.jit.trace(torch.rand(3, 4), check_inputs=[(torch.rand(3, 4),)])(foo)
```
```
torch.jit.TracingCheckError: Tracing failed sanity checks!
ERROR: Traced function outputs do not match the Python function outputs.
ERROR: Graphs differed across invocations!
Graph diff:
graph(%0 : Dynamic) {
- %1 : Dynamic = aten::neg(%0)
+ %1 : Dynamic = prim::Constant[value=<Tensor>]()
%2 : int = prim::Constant[value=1]()
%3 : Dynamic = aten::add(%0, %1, %2)
return (%3);
}
Node diff:
- %1 : Dynamic = aten::neg(%0)
+ %1 : Dynamic = prim::Constant[value=<Tensor>]()
Trace source location:
test.py(5): foo
/Users/jamesreed/onnx-fairseq/pytorch/torch/jit/__init__.py(402): wrapper
test.py(8): <module>
Check source location:
test.py(6): foo
/Users/jamesreed/onnx-fairseq/pytorch/torch/jit/__init__.py(281): check_trace
/Users/jamesreed/onnx-fairseq/pytorch/torch/jit/__init__.py(408): wrapper
test.py(8): <module>
```
The following two examples show instances where program semantics are lost in the Python -> trace transformation, and repeated invocation does not give us useful debug information. Further design in underway for catching these scenarios.
```
import torch
torch.jit.trace(torch.rand(3, 4), check_inputs=[(torch.rand(3, 4),)])
def foo(x):
for i in range(3):
x[i, :] = torch.zeros(4)
return x
```
```
torch.jit.TracingCheckError: Tracing failed sanity checks!
ERROR: Traced function outputs do not match the Python function outputs.
Exception:
Not equal to tolerance rtol=1e-07, atol=0
(mismatch 100.0%)
x: array([0.830221, 0.915481, 0.940281, 0.555241], dtype=float32)
y: array([0., 0., 0., 0.], dtype=float32)
```
==
```
import torch
torch.jit.trace(torch.rand(3, 4), check_inputs=[(torch.rand(5, 6),)])
def foo(x):
x.view(-1).add_(-x.view(-1))
return x
```
```
torch.jit.TracingCheckError: Tracing failed sanity checks!
ERROR: Traced function outputs do not match the Python function outputs.
Exception:
Not equal to tolerance rtol=1e-07, atol=0
(mismatch 100.0%)
x: array([0.734441, 0.445327, 0.640592, 0.30076 , 0.891674, 0.124771],
dtype=float32)
y: array([0., 0., 0., 0., 0., 0.], dtype=float32)
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/10841
Differential Revision: D9499945
Pulled By: jamesr66a
fbshipit-source-id: 1f842a32d0b0645259cc43b29700b86d99c59a45
Summary:
* Fix the necessary pathways so that tuples and lists can be inputs to the script.
* prevent linear algebra functions from being run in shape prop because
they frequently will error out for nonsense data.
* favor schema-driven python input conversion where possible.
remaining cases where we directly create Stacks without schema are
only for debugging
* Make the error messages when calling script/trace functions more pythonic
* Simplify FlattenTuples -- now that tuples are supported we can choose to only flatten tuples when needed. This may have to be revisited pending onnx test results, but is necessary for making tuple io work.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/10812
Differential Revision: D9477982
Pulled By: zdevito
fbshipit-source-id: ed06fc426e6ef6deb404602a26c435a7fc40ea0c
Summary:
Please review the expects carefully to make sure there are no regressions. I tried to go over them one by one when they changed, but it's sometimes easy to miss finer details.
Summary of changes:
- Renamed `TensorType` to `CompleteTensorType`. Added a new `TensorType` which records only the scalar type, number of dimensions, and device of a value. The argument behind the rename is to encourage people to use `CompleteTensorType` less, as most passes will only have limited information available. To make transition easier `complete_type->cast<TensorType>()` works, and makes our passes work with both kinds of specialization if they don't need extra the extra detail.
- Renamed `ArgumentSpec` to `CompleteArgumentSpec`. Added a new `ArgumentSpec`, which matches argument only at the level of the new `TensorType`.
- Shape analysis can process graphs with both `CompleteTensorType` and `TensorType`.
- Fuser was a part that heavily relied on full shape information being available. Now, we simply try to fuse the largest possible graphs, and have to do run-time checks to make sure they match the code we generate. If they don't, we fall back to regular interpretation. The shape checks are implementing using an optimized method exploiting algebraic properties of shapes with broadcasting, and the relations of broadcasting with pointwise ops. A full written proof of correctness of the shape checking algorithm is included in a comment in `graph_fuser.cpp`.
zdevito ezyang mruberry ngimel csarofeen
Pull Request resolved: https://github.com/pytorch/pytorch/pull/10844
Differential Revision: D9498705
Pulled By: apaszke
fbshipit-source-id: 0c53c2fcebd871cc2a29c260f8d012276479cc61
Summary:
When matching schema, first try to match without adding TensorToNum conversions. Then make another pass where TensorToNum conversions are allowed.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/10180
Differential Revision: D9438153
Pulled By: eellison
fbshipit-source-id: 80541b5abd06e9d4187e89dda751f44dab6f58c5
Summary:
Part of #10774.
This PR does the following:
- Support ast.ExtSlice in the frontend. This is done by returning a
list of ast.Index and ast.Slice.
- Support multidimensional indexing with ints and slices
The general approach is to desugar multidimensional indexing into
at::slice, at::select operations. This is exactly how normal pytorch
does indexing (by desugaring it into at::slice, at::select, and other ops).
I used [this code](https://github.com/pytorch/pytorch/blob/master/torch/csrc/autograd/python_variable_indexing.cpp) as reference.
We should be able to copy the rest of this to implement the missing
indexing features in script (indexing with ellipses, tensors, sequences, etc).
After I'm done implementing the missing indexing features in future prs, I can try to
templatize python_variable_indexing.cpp so that it can work with both JIT
script and normal pytorch indexing, but right now I'm not sure if that's
a good idea or not.
cc zdevito jamesr66a apaszke wanchaol
Pull Request resolved: https://github.com/pytorch/pytorch/pull/10787
Differential Revision: D9481402
Pulled By: zou3519
fbshipit-source-id: 78c9fa42771a037d157879e23e20b87401cf1837
Summary:
Things like `zeros(1,2,3, dtype=torch.int)` are now supported in the script by altering tryMatchSchema to auto-construct the list `[1,2,3]` when it sees inlined members of the list as the last positional arguments.
I suggest reading the commits individually, since the first two incrementally change how we do tryMatchSchema to get it ready for adding vararg list conversion, while the third actually does the modification.
closes#10632closes#8516
Pull Request resolved: https://github.com/pytorch/pytorch/pull/10250
Differential Revision: D9478235
Pulled By: zdevito
fbshipit-source-id: 0c48caf7a6184e463d9293d97015e9884758ef9c
Summary:
When emitting if Branches, check that the types on each value returned are equivalent. As with reassignment of values, tensors are not forced to be the same shape or subtype.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/10281
Differential Revision: D9466566
Pulled By: eellison
fbshipit-source-id: 746abdeb34a0f68806b8e73726ad5003b536911c
Summary:
Augassign (i.e., `x += 1`) gets desugared to an assignment of a binop (`x = x + 1`).
Right now we assert that the RHS of the binop is a tensor,
but it really doesn't have to be because we support scalar/scalar ops and also
list-list ops (i.e., `[1, 2] + [2, 3]`).
Pull Request resolved: https://github.com/pytorch/pytorch/pull/10730
Differential Revision: D9465110
Pulled By: zou3519
fbshipit-source-id: 7b118622701f09ce356aca81b8db743d9611097b
Summary:
This PR adds support for using custom ops in ScriptModules, the last step for our custom op strategy. You can now write
```
import torch
torch.ops.load_library('libcustom_ops.so')
class Model(torch.jit.ScriptModule):
def __init__(self):
super(Model, self).__init__()
torch.jit.script_method
def forward(self, input):
return torch.ops.custom.op(input) + 1
model = Model()
model.forward(torch.ones(5)) # Works
model.save("model.pt") # Works
model = torch.jit.load("model.pt") # Works
```
You can then load the `model.pt` in C++ and execute its `forward` method!
Missing for this was the fact that the script compiler didn't know to convert `ops.custom.op` into a `BuiltinFunction` which then emits a function call. For this I came up with the following strategy inside `torch/csrc/jit/scrip/init.cpp`:
1. When we access `torch.ops`, we return a `CustomOpValue` (subclass of `PythonValue`), whose purpose is only to return a `CustomOpNamespaceValue` (subclass of `PythonValue`) whenever something under it is accessed.
2. `CustomOpNamespaceValue` will then for each field accessed on it return a `BuiltinFunction`.
This doesn't reduce performance for any calls that are not to `torch.ops` (as opposed to inspecting every function call's name the call site, for example).
I also had to fix `BuiltinFunction` to not assume the namespace is always `aten::`.
A lot of other changes are just tidying up the Python and C++ test harness before I integrate it in CI.
zdevito dzhulgakov
Pull Request resolved: https://github.com/pytorch/pytorch/pull/10610
Differential Revision: D9387832
Pulled By: goldsborough
fbshipit-source-id: c00f431db56c7502a66fe1f813fe78067f428ecb
Summary:
This will make the common case more natural (no need to do `_construct_empty_tensor_list()`)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/10705
Differential Revision: D9411622
Pulled By: michaelsuo
fbshipit-source-id: 2d91fbc5787426748d6e1c8e7bbeee737544dc96
Summary:
After this, all combinations of {String frontend, Python AST Frontend}{Python 3-style type annotations, MyPy-style type comments}{Script method, Script function} should properly accept type annotations.
Possible TODOs:
- Clean up the functions marked HACK
- Clean up the Subscript tree-view to better match the Python AST versions
- Can we use this for Python functions? That's the only place annotations.get_signature() is still needed
Pull Request resolved: https://github.com/pytorch/pytorch/pull/10279
Differential Revision: D9319726
Pulled By: jamesr66a
fbshipit-source-id: b13f7d4f066b0283d4fc1421a1abb9305c3b28fa
Summary:
This commit adds the ability to insert a node with inputs, using the schema to check the inputs are valid types, fill in any default values, and perform standard implicit conversions. Since it is schema based, it will discover and use the right overload.
Constructors to `NamedValue` enable it to be constructed using `IValue` constants so it is possible to use constant values in the input list as well:
```
g.insert(aten::add, {v, 3});
```
Keyword arguments are also supported:
```
g.insert(aten::add, {v}, {{"other", t}, {"scalar", 1}});
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/10198
Differential Revision: D9307252
Pulled By: zdevito
fbshipit-source-id: 644620aa85047d1eae1288383a619d50fec44d9b
Summary:
Previously, `tensor[i:]` was transformed to `tensor[i:-1]`. This incorrectly leaves off the last element. Noticed this when implementing slicing for list types.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/10286
Differential Revision: D9193292
Pulled By: michaelsuo
fbshipit-source-id: df372b815f9a3b8029830dd9e8769f9985a890e7
Summary:
I changed the name of this builtin to match Python's native style, but forgot to change the compiler error to match.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/10265
Differential Revision: D9192963
Pulled By: michaelsuo
fbshipit-source-id: 225ca4cd50fbbe3b31c369deeb3123a84342aab1
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/10130
Update some include paths to make them internally consistent
Reviewed By: ezyang
Differential Revision: D9119906
fbshipit-source-id: b44e5cab8e8e795ee18afe9ffc6caf1f2b413467
Summary:
This PR adds a way to infer the JIT/script schema of a function from its signature, and then create an operator from the schema and implementation. The implementation function is wrapped into another function, which pops values from the stack into an argument tuple, then invokes the function and pushes the return value back onto the stack, sometimes unpacking the return value if it is a tuple.
Currently the method is called `createOperator`. We may want to think of a nicer way of registering ops in tandem with `RegisterOperators`. It might be very cumbersome to add a template constructor to `Operator`, so maybe we can come up with a chaining method on `RegisterOperators` like `RegisterOperators(schema, func).op(schema.func).op(schema, func)` -- it has to work at startup time (for a static variable) though. We can solve this in another PR.
zdevito apaszke smessmer dzhulgakov
Pull Request resolved: https://github.com/pytorch/pytorch/pull/10048
Differential Revision: D9125975
Pulled By: goldsborough
fbshipit-source-id: de9e59888757573284a43787ae5d94384bfe8f9a
Summary:
The PR allows int→float and float→int casts. Current we only allow `tensor→int` and `tensor→float` casts.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/10168
Differential Revision: D9141163
Pulled By: wanchaol
fbshipit-source-id: 5e5591a98b4985a675641dfc9a385b2a0bf8e208
Summary:
Previously, `foo = [bar, baz]` would construct a TupleType of fixed arity. This would cause code like:
```
foo = [2]
if True:
foo = [2, 2]
```
to fail to compile, since `(int)` is not the same as `(int, int)`.
This PR changes things so that list literals construct ListTypes, which can be resized.
Potentially breaking changes introduced:
- Empty list literals are now disallowed, `_constructEmptyFooList()` builtins are required to replace them.
- Iterable variable unpacking where the rhs is a list is now disallowed. (Tuples still work)
- Lists must have a single type.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/10193
Differential Revision: D9147166
Pulled By: michaelsuo
fbshipit-source-id: bbd1b97b0b6b7cb0e6f9d6aefa1ee9c731e63039
Summary:
* Changes `insertConstant(g, val)` to `g.insertConstant(val)`.
* Moves SourceRange to its own file to enable it.
* Cleans up dead attribute code in schema matching and graph.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/10177
Differential Revision: D9137789
Pulled By: zdevito
fbshipit-source-id: 8a73cfb01a576f02e7e4dce019be9c0a0002989d
Summary:
This PR adds strings to the ast and implements them for print statements. Strings are lifted as attributes to the print node. They must be arguments to print itself, not as an argument for an object that is passed to print. If they are encountered elsewhere a NYI exception will be thrown.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/9324
Reviewed By: jramseyer
Differential Revision: D8807128
Pulled By: eellison
fbshipit-source-id: 984401ff458ed18d473c6d1bd86750e56c77d078
Summary:
Previously, the parser was emitting list literals for tuples, but the IR was representing list literals internally with TupleTypes.
For implementing most list operations, I think it will be helpful distinguish between lists (dynamic size, homogeneous types) and tuples (fixed arity, heterogeneous types)
This diff modifies the parser logic to emit tuple literals. This frees us to represent lists as ListType in the IR, while still properly mapping tuple literals to TupleTypes.
A following diff will actually switch over list literals to emit ListTypes.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/10128
Differential Revision: D9121305
Pulled By: michaelsuo
fbshipit-source-id: e0cad07ae8bac680f7f8113d10e5129d5a1a511d
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/9740
- Remove implicit ArrayRef -> vector conversion
- Fix 4 call sites that accidentally did an implicit expensive vector conversion but wouldn't have needed to
- Remove explicit vector conversion from 4 call sites that also didn't need to do that
Reviewed By: ezyang
Differential Revision: D8961693
fbshipit-source-id: 980da9f988083c0072497f9dbcbbf6f516fa311c
Summary:
This should prevent slow startup times, and will not report as many
errors during static initialization time which are hard to debug
ezyang apaszke
Pull Request resolved: https://github.com/pytorch/pytorch/pull/9801
Reviewed By: goldsborough
Differential Revision: D8986603
Pulled By: zdevito
fbshipit-source-id: 440d43ab5e8cffe0b15118cb5fda36391ed06dbc
Summary:
This lays out initial support for taking and returning a richer set
of types than only tensors. Floats and ints are already valid, lists are
straightforward to add, tuples need some discussion.
Based on top of #9948. Review only the last commit.
zdevito
Pull Request resolved: https://github.com/pytorch/pytorch/pull/9969
Reviewed By: zdevito
Differential Revision: D9076973
Pulled By: apaszke
fbshipit-source-id: 5a1fe912ea6b79ab2bfd0dcce265eb05855b5ff0
Summary:
Supersedes #8925
This PR fixes#8502, it fixes the gradients problem for clamp when passing None to the function, and add support for the NoneLiteral and NoneType in script to enable clamp tests. Now we could have corner cases like:
```python
torch.jit.script
def func():
x = torch.randn(3, 3, requires_grad=True)
y = torch.clamp(x, None, 0) # max = 0
y = torch.clamp(x, min=None, max=0)
```
In both JIT and Aten, we use Scalar(NAN) as a sentinel value when passing None type to function clamp, this is the current way we used to support None type in JIT and to solve the gradient problem when user explicitly passing None into clamp.
In JIT side, we create a tensor(NAN) and undefinedTensor if we encounter None when matching the function schema, and later in the interpreter, it will translate to Scalar(NAN) if needed.
Ideally we don't need clamp_min and clamp_max in ATenNative/Autograd and could only support clamp after this change, but since bunch of other operators (e.g. Activation.cpp, Loss.cpp) is using clamp_min in several places, we will still have the functions available, but all python invocations will only call clamp instead of clamp_min/max (with calling underlying th_max/th_min in clamp).
zdevito jamesr66a
Pull Request resolved: https://github.com/pytorch/pytorch/pull/9596
Reviewed By: zdevito
Differential Revision: D8940839
Pulled By: wanchaol
fbshipit-source-id: c543a867b82e0ab8c99384773b173fdde2605d28
Summary:
Based on top of #9763 (first 3 commits belong to that PR). The first commits from this PR are "Stop using attributes ..."
I tried to separate the changes into fairly meaningful commits. I can't split them up into smaller PRs, because everything starts working and all tests pass only after the whole sequence, but hopefully this will make reviewing somewhat easier.
Known issues/regressions/future tasks:
- `aten::lerp` and `aten::clamp` are no longer fusable
- `CreateAutodiffSubgraphs` needs a rewrite
- It is much more strict now, and will miss a lot of opportunities, especially when viewing ops are involved. Our previous approach was "ignore the assumption on shape availability in gradient formulas to determine differentiability, and hope that shape prop will be robust enough to actually deliver them before we differentiate", which obviously doesn't scale well to more complex cases. We should either work on reducing the size dependency of grad formulas (feasible e.g. for `view`/`reshape`, unfeasible for `squeeze`/`unsqueeze`), or make `CreateAutodiffSubgraphs` integrate some kind of "I could integrate this node into an AD subgraph, but will I be able to infer the shape of its input" reasoning (kind of like a limited shape prop, that doesn't infer anything, and only tells if it *could* infer something).
- It sometimes creates constant-only (or constants + one node) graphs, which is useless
- Broken `aten::add` in auto-batching, because it gained a non-tensor input. I changed the test for pointwise operations to use `aten::mul` instead, but I needed to disable the LSTM cell test. I'm not sure how scalar constants should be implemented in this case, because I don't fully understand our format. cc: ChunliF
- Graph import does some hacks to recover type of constants. This code should be removed once we'll gain the ability to export the IR along with value types.
- There's still a fair amount of dead code that can be removed. I didn't want to make this diff any bigger, and removing it is an easy task.
- Graph fuser could be improved to use signature matching (possibly using `OperatorSet`) instead of basing on node kinds.
- Manual constant propagation for the `ListConstruct` node in `torch/onnx/utils.py` should be replaced with a proper constant propagation pass (or we should ensure that the one we have handles at least this case before we remove this code).
zdevito
Pull Request resolved: https://github.com/pytorch/pytorch/pull/9807
Reviewed By: ezyang
Differential Revision: D9004285
Pulled By: apaszke
fbshipit-source-id: fe88026a765f6b687354add034c86402362508b7
Summary:
Follow up task of #9584.
Commit 1:
- change expect/cast to return shared pointers instead of raw pointer
- isSubtypeOf accept TypePtr instead. Use `x->isSubtypeOf(NumberType::get())` rather than `x->isSubtypeOf(*NumberType::get())`
Commit 2:
- to address enable_shared_from_this pitfalls, we make the constructor private and expose the factory method to make sure user can only create it using our factory method.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/9786
Reviewed By: zdevito
Differential Revision: D8980441
Pulled By: wanchaol
fbshipit-source-id: e5c923fc57a701014310e77cf29985b43bb25364
Summary:
I got some tensor->variable conversion exceptions from `torch/csrc/autograd/variable.h`, which used the `TORCH_ASSERTM` macros instead of `AT_CHECK`, so they didn't have backtraces. This was such a substantial loss for debugability that I decided to update the whole codebase to use the backtrace-enabled ATen macros instead of `TORCH_ASSERT` and `JIT_ASSERT`, the latter having been an alias of the former.
ezyang apaszke zdevito
Pull Request resolved: https://github.com/pytorch/pytorch/pull/9575
Differential Revision: D8924566
Pulled By: goldsborough
fbshipit-source-id: 7a4013b13eec9dbf024cef94cf49fca72f61d441
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/9718
This patch switches the interpreter to use IValue's primitive numbers rather than tensors for computing on integers and floats. In addition to preparing the interpreter for first-class support of other types, this cleans up the handling of primitive numbers, making it possible to just use the normal operator overloading dispatch to find the right implementation for numbers. As a result of this change, a lot of other functionality needed to be updated since it was the first time we use non-tensors in a lot of places in the code base.
Notes:
* Fixes code_template.py so that multi-line strings are indented correctly when used on a standalone line
* Cast operators (`int(x)`) now are functional. Some tests have addition conversions to integers because
we no longer allow implicit tensor -> integer conversions following the same convention as in python
* prim::ListConstruct/createList has been added to the interpreter for creating lists and this has
replaced aten::stack for integers lists
* gen_jit_dispatch.py has been refactored so that non-tensor types use operators on IValues to extract
the primitives
* IValue gains a .to<T> method that is the equivalent of tensor_as but for IValue instead of at::Tensor
* `constant_as<T>` is switched over to using IValues's `.to<T>` method, to make conversion from constant->IValue->C++ type
more consistent. This functionality combined with `toIValue(Value*)` replaces the `tensor_as` and `as_tensor` family of functions.
* conditional expressions (if, loop) and operators related to them are now computed on integers rather than tensors
* IValue gains constructors for constructing from at::Scalar and converting to it. However, IValue itself will always store
the scalars as a double or int64.
* To align with python 3 syntax, TK_INT, TK_FLOAT, and TK_BOOL have been removed from the parser, and int/float/bool are just treated as special identifiers in the compiler,
along with print. These are represented as special sugared values with a `call` method implemented. For int/float/bool this implements casting behavior.
* Dropped shared_from_this from Type/Module. They were not needed and they making debugging harder because they internally throw/catch exceptions.
* Shape propagation has been updated to support running nodes that include floating point primitive types, this required some refactoring of internal functions.
* TensorToNum and NumToTensor have actual implementations as operators now
* regster_prim_ops now contains implementations of math operators for float/int primitive types, and for mixed (prim <+> tensor) versions. This removes the need for special handling in compiler.cpp
* Primitive math is now entirely handled by letting the compiler choose the right overloads. This removes tons of special casing in the compiler.
* incorporates eellison's change to allow casting from return values. Due to the addition of primitive support, the code need slight modifications, so I just pre-merged it here.
* stack.h gains generic vararg versions of push/pop that know how to convert to/from C++ types:
```
at::Tensor a;
at::Scalar b;
pop(stack, a, b);
at::Tensor c = a + b;
push(stack, c);
```
apaszke
Pull Request resolved: https://github.com/pytorch/pytorch/pull/9584
Reviewed By: apaszke
Differential Revision: D8910546
Pulled By: zdevito
fbshipit-source-id: 0f3e60d4d22217f196a8f606549430e43b7e7e30
Summary:
IValue is short for interpreter value. It is used frequently so a short name is important.
This will allow us to implement more non-tensor types in an efficient way and remove
many hacks from the compiler.
This PR is limited. It only introduces IValue and changes interpreter to use it.
Follow up PRs will:
* Change the way aten_ops consume non-tensor types so that integer lists,
are no longer represented as Tensors.
* Introduce TensorList as a fundamental type and remove all vararg handling in gen_jit_dispatch
* Change the compiler to implement math on primitive numbers rather than converting to tensors.
jamesr66a apaszke
Pull Request resolved: https://github.com/pytorch/pytorch/pull/9368
Reviewed By: ezyang
Differential Revision: D8817598
Pulled By: zdevito
fbshipit-source-id: 29dce80611ce5f6384234de9d12a67861d2b112f
Summary:
This is a series of two commits that should probably be read separately. They are stacked on top of #9018 since the second commit requires it for correctness.
Commit 1
=======
This commit is the first in a series that will clean up how we handle declaring operators and intrinsics in the JIT to make it more modular and readable. This introduces readable declarations that can be used to register operators and switches gen_jit_dispatch to generate this schema. A follow up PR will remove the dispatch keys like "add-3" and resolve ops directly based on the registered schema, further simplifying the generation process.
* Switches schema over to parsed declarations, in the future this will allow something like:
```
registry.register_intrinsic("foo(Tensor a, Tensor b) -> Tensor", [](Stack& stack) {
...
})
```
This will allow the scalable registration of intrinsics for lists, tuples, and other ops, as long as meta-data for these ops (e.g. derivatives and size propagation routines).
The declarations resemble those used by PythonArgParser but have been singificantly cleaned up to minimize the number of types that can appear in the declaration. We should strive to get the other parts of PyTorch switched over to this restricted declaration set when possible, but it is too much to do in a single PR. My hope is that eventually we will use a very similar language to describe declarations in C10, and this can serve as a guide for that.
Parsing is done using the script lexer, so it is very robust to whitespace and extensible for future types.
This removes the other way we encoded schema, and makes it easier to see what schema are registered.
Current generated declarations: https://gist.github.com/zdevito/a96a17766fb3a098d69a91ee00abaaf6
* Switches how we handle attempting to use an integer in the place of a fixed-sized int list, such as in conv (e.g. 'int[3] stride=1'). Now that we can statically distinguish between int and Tensor, we handle the expansion as an implicit conversion in the compiler. This allows us to simplify the interpreter since it no longer needs to handle the conversion itself.
* Schema declarations have been changed so that they match the type system in the IR exactly. In particular, attribute_info which was used by liftConstantAttributes has been dropped and constant attributes are lifted purely based on the type of the input. Type conversions in compiler have been simplified due to this change.
* Error highlighting in ErrorReport now only reports at most 20 lines of code, to make reading where an error occurred easier.
Commit 2
=======
This commit unifies aten_dispatch and aten_schema into a single Operator object that both contains schema and implementation information. In the future we can use this object to also contain functionality like shape prop and autodiff needed by all operators. Operators are registered globally, and dispatch logic uses the schema information to figure out which variant to use. Descriptor keys, a frequent source of inscrutable debug errors, have been removed.
* Introduce Operator, to replace TensorOp. Unlike TensorOp, we use Operator for all op implementations, including primitives that may occur in the graphs. The only exceptions are ops that are only known to the interpreter like jumps, and GraphExecutors where we need to record additional debug info.
* Adds a global registry for Operator implementations. aten_dispatch.cpp turns into register_aten_ops.cpp, which registers all the Operators for aten with the operator registry. register_prim_ops.cpp now contains the implementations for primitive operators that used to be in the interpreter. This means that it is now safe to use `getOperation(node)` to lookup the true interpreter function for the node, which will simplify const-propagation passes.
* Remove addInterpreterOpHandler in favor of global operator registry.
* Instead of descriptors, we match Node arguments directly against FunctionSchema describing expected inputs in `matchSchema`. `matchSchema` knows how parse both attributes and positional inputs from a node and match it to the appropriate registered operator. Debug error messages when we try to run an invalid operator are significantly improved: they now automatically display the schema for the op with the same name that are registered.
* Merge aten_schema into regsiter_aten_ops. Each Operator takes a string schema which is parsed to determine when to dispatch to that op.
* Cleans up gen_jit_dispatch.py now that we do not need to write out descriptors. In particular, skip_scalar_overloads can be removed since Richard's code sorts declarations to put Tensor, Tensor declarations first.
* remove matchSchemaAndLiftConstantAttributes and use emitBuiltinCall instead to remove code duplication
* refactor stack manipulation functions into a separate header file.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/8885
Reviewed By: jamesr66a
Differential Revision: D8751048
Pulled By: zdevito
fbshipit-source-id: 312aabfbf88307c5f6ab947b6caf691468b94557
Summary:
Here's an improved error message. Let me know if this change makes the errors a little clearer.
Closes https://github.com/pytorch/pytorch/pull/9212
Reviewed By: soumith
Differential Revision: D8752896
Pulled By: jramseyer
fbshipit-source-id: d2bd8462c3ddf14acd3de56a4c1aeb75a9bc4067
Addresses #8177
A design doc can be found here: [gist](https://gist.github.com/zou3519/4b7f13f03cc9f3612bd9363e6405fa0a) version or [quip](https://fb.quip.com/azL1AqUckBdo) version
General approach:
- Add NumberType, FloatType, IntType to represent Python numbers, floats and ints.
- Emit these types for python literals
- Change aten_schema such that Scalars are NumberType, int64_t and bool are IntType.
- Emit aten::type_as, prim::NumToTensor, and prim::TensorToNum nodes for tensor-number math. (see examples below)
- Erase NumberType, prim::NumToTensor, and prim::TensorToNum for ONNX export
### Tensor/number math
```
import torch
@torch.jit.script
def fn(x):
return x + 1
```
```
graph(%x : Dynamic) {
%1 : int = prim::Constant[value={1}]()
%2 : Dynamic = prim::NumToTensor(%1)
%3 : Dynamic = aten::type_as(%2, %x)
%4 : Dynamic = aten::add[alpha={1}](%x, %4)
return (%5);
}
```
### Number/Number Math
```
import torch
@torch.jit.script
def fn(zero):
c = 1 + 1
return zero + c
```
```
graph(%zero : Dynamic) {
%1 : int = prim::Constant[value={1}]()
%2 : int = prim::Constant[value={1}]()
%3 : Dynamic = prim::num_to_tensor(%1)
%4 : Dynamic = prim::num_to_tensor(%2)
%5 : Dynamic = aten::add[alpha={1}](%3, %4)
%c : int = prim::TensorToNum(%6) # this is the result of the addition
...
return (%13);
}
```
List of squashed commits:
* Introduce Python Number types
Added: IntType, FloatType, NumberType with
IntType <: NumberType
FloatType <: NumberType
Changed aten_schema so arguments have corresponding types
* Emit a NumberType for python literals.
Also emit a NumberType for Scalar default values.
* Add prim::NumToTensor and prim::TensorToNum
* Add DynamicType -> NumberType implicit cast for bc
* Better ensureTensor error message
* Add ensureTensorOrNumber. Allow passing Number to some functions
Like the range() construct and slices
* Patch IntList to work.
IntList is still a DynamicType in the frontend: a tensor gets built from
a List[int].
Also, IntList[1] is a "union between int and IntList" the way it is
implemented. If the frontend sees an int being passed for an IntList[1]
arg, it converts it to a tensor as well.
* Enforce some order on schemas to avoid overload ambiguity
add(Tensor, Tensor) should appear earlier than add(Tensor, Scalar). This
matches the order in which python_arg_parser parses its arguments.
* Disable std_dim and var_dim tests.
With the new schema information, std(input, keepdim) and std(input, dim)
are ambiguous. This will need to be fixed at a later date.
* Add NumberType erasure pass.
This is used for ONNX export and to ensure that NumberType information
doesn't reach the interpreter
* Add support for mixed tensor/number math ops.
* Tests for new functionality.
Includes:
- Tensor/number math
- number/number math
- EraseNumberTypes pass test
* Patch tests
Update expect tests for:
- decompose_addmm
- loop unrolling tests
Because python numbers are now NumberType, they cannot be returned by
functions anymore. Work around this by using "torch.full", or by adding
a tensor([0]) (taken from FIXME_zerol()). Both approaches are used
because torch.full is more readable, but it is broken in some cases.
* Add erase_number_types to torch/CMakeLists.txt
* Move math back to emitSimpleExpr from emitSugaredExpr
* Remove some dead lines
* Renable some excluded script/trace tests that are fixed.
* Move some tests to expected failure
* Address some comments (more addressing to come)
* Erase relevant aten::type_as nodes in EraseNumberTypes
I also changed it so that EraseNumberTypes is only called for ONNX
export. It is no longer used to prevent
prim::NumToTensor/prim::TensorToNum from reaching shape_analysis or
interpreter.cpp.
shape_analysis infers the type of the output of these nodes to be the
same as their input.
intepreter.cpp treats both of these nodes as no-ops.
* Add reminder to fix std/var
* Call EraseNumberTypes only when exporting a script module
* Update expects after rebase
* enable captured inputs for if Stmt to fix the carried deps bug in nested
blocks
* postpone captured inputs deletion and add new test case
* recursively generate captured values for nested loops
* check asSimple when recursively create captured input
