Use conditional imports: when running under dynamo, import the original NumPy not torch._numpy. This is what we want to trace, not our implementation.
With this, the test suite passes with and without `PYTORCH_TEST_WITH_DYNAMO=1` (modulo a couple of test modules which are not meant to be compiled, e.g. `test_nep50_examples`). There are two new decorators, `x{fail,pass}ifTorchDynamo`, the `xpass` in most cases indicates a graph break and a fallback to eager for things we do not implement.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/110401
Approved by: https://github.com/lezcano
AOTAutograd's handling for resize_() isn't fully robust (and on top of that, functionalization can potentially give up and raise an error if the tensor you're resizing has outstanding views).
So given that, and given that resize_() is rare, I updated dynamo to graph break on resize_() instead.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/111553
Approved by: https://github.com/ezyang
Did some easy fixes from enabling TRY200. Most of these seem like oversights instead of intentional. The proper way to silence intentional errors is with `from None` to note that you thought about whether it should contain the cause and decided against it.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/111496
Approved by: https://github.com/malfet
Fixes part 1 of https://github.com/pytorch/pytorch/issues/111370#issuecomment-1764730773
While at it, add a test for numpy ndarray `.size` attribute. This started as an attempt to remove the delegation of what looks like a `.size()` method --- which does not exist in numpy --- on the same line this patch adds a `tolist` to.
But this is apparently needed for something else and existing tests start failing. Thus, declare it as _ain't broken don't fix_, and only keep the test. Can remove the test if wanted though.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/111382
Approved by: https://github.com/lezcano
Fixes#109604
Resubmit gh-109715 + several skips and small fixes to make tests pass.
The main fix here is by @ysiraichi : previously, dynamo did not resume tracing numpy ndarrays after a graph break.
While at it, fix several small issues Yukio's fix uncovers:
- graph break gracefully on numpy dtypes which do not map to torch.dtypes (uint16 etc)
- recognize array scalars in dynamo, treat them as 0D ndarrays
- make sure that iterating over torch.ndarray generates arrays not bare tensors
Pull Request resolved: https://github.com/pytorch/pytorch/pull/110512
Approved by: https://github.com/lezcano
The main thrust of the initial effort here was to capture `register_hook` calls on tensors in compile regions. The first part of this was done in https://github.com/pytorch/pytorch/pull/108903 wherein we added support for register_hook input tensors.
The distinction between input and intermediary is due to implementation differences.
There are 2 kinds of hooks:
1) Hooks on objects with sources (inputs, params)
2) Hooks on objects w/o sources (intermediaries, and outputs).
Note: As outputs can be made simple by how dynamo handles residuals, they could actually be handled as if they were inputs, but, for the sake of this PR, we will refer to hooks as either hooks on inputs (sourced), or hooks on intermediaries (not sourced).
**The plan:**
For tensors w/ a source: (The PR above)
We record registered hooks, store them as a global, and associate them with the tensor in residuals. This means that when dynamo goes to create the frame, where we produce bytecode to stitch together our PT2 modified bytecode with the original eager code, we call register_hook. This registration of hooks in residuals is sound because (a) it happens right after a Pt2 frame region ends and (b) we know that the tensor is alive in f_locals, f_globals, or a module in the users invoking frame. This means we can soundly know it will be around to invoke register_hook on. As long as we guard on the identity of the lifted function, this is sound to do.
For tensors w/o a source: (This PR)
Ostensibly, the most correct and complete solution would be to smuggle hooks into a runtime wrapper in aot_autograd, where all the items the hooks close over are lifted to inputs as necessary and passed alongside the user provided function. This is necessary so that we can properly trace out and capture all the mutations within the user defined hook at backwards time.
This is too complicated - so, we limited the scope of this initial PR to a simple subset of hooks:
- Hooks must have a source (be known to us already, not a lambda or intermediary defined function)
- We must be tracing under compiled autograd
**The flow**:
We use the HOP added in https://github.com/pytorch/pytorch/pull/109690/files, referred to as the HOP below.
1) We intercept register_hook calls and wrap the user defined fn in the HOP
2) We write a `_register_hook_trampoline` to the graph that is a local no-arg function that is invoked as a call_function in the dynamo graph
3) aot_autograd inlines through it during its trace, and sees the HOP
4) the HOP preserves itself in the graph - it does not get traced into
5) During backwards, compiled_autograd installs the HOP under a hook call
6) When compiled_autograd enters compilation over its generated graph, dynamo traces the contents of the hook
Pull Request resolved: https://github.com/pytorch/pytorch/pull/109537
Approved by: https://github.com/ezyang
The strategy in this PR is pretty straightforward.
There are 2 kinds of hooks:
1) Hooks on objects with sources (inputs, params)
2) Hooks on objects w/o sources (intermediaries, and outputs).
Note: As outputs can be made simple by how dynamo handles residuals, they could actually be handled as if they were inputs, but, for the sake of this PR, we will refer to hooks as either hooks on inputs (sourced), or hooks on intermediaries (not sourced).
The plan:
**For tensors w/ a source:**
We record registered hooks, store them as a global, and associate them with the tensor in residuals. This means that when dynamo goes to create the frame, where we produce bytecode to stitch together our PT2 modified bytecode with the original eager code, we call `register_hook`. This registration of hooks in residuals is sound because (a) it happens right after a Pt2 frame region ends and (b) we know that the tensor is alive in f_locals, f_globals, or a module in the users invoking frame. This means we can soundly know it will be around to invoke `register_hook` on. As long as we guard on the identity of the lifted function, this is sound to do.
**For tensors w/o a source:**
Graph break - we will support this in a subsequent PR
**Handles:**
An interesting new component here is the creation of a `STORE_FAST `->`LOAD_FAST` associated with the handle, the return result of `register_hook`. If the user code stored the result of `register_hook` in a handle, we need to honor that. We do so by interceding into `STORE_FAST`, and recording the name of the local variable as directed by user code. We then honor that same name in the reconstructed bytecode. If the user did not store a hook, we merely pop the produced value to preserve the stack.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/108903
Approved by: https://github.com/ezyang
ghstack dependencies: #108846, #109092
RFC: https://github.com/pytorch/rfcs/pull/54
First commit is the contents of https://github.com/Quansight-Labs/numpy_pytorch_interop/
We have already been using this in core for the last few months as a external dependency. This PR pulls all these into core.
In the next commits, I do a number of things in this order
- Fix a few small issues
- Make the tests that this PR adds pass
- Bend backwards until lintrunner passes
- Remove the optional dependency on `torch_np` and simply rely on the upstreamed code
- Fix a number dynamo tests that were passing before (they were not tasting anything I think) and are not passing now.
Missing from this PR (but not blocking):
- Have a flag that deactivates tracing NumPy functions and simply breaks. There used to be one but after the merge stopped working and I removed it. @lezcano to investigate.
- https://github.com/pytorch/pytorch/pull/106431#issuecomment-1667079543. @voznesenskym to submit a fix after we merge.
All the tests in `tests/torch_np` take about 75s to run.
This was a work by @ev-br, @rgommers @honno and I. I did not create this PR via ghstack (which would have been convenient) as this is a collaboration, and ghstack doesn't allow for shared contributions.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/106211
Approved by: https://github.com/ezyang
This PR adds initial dynamo support for DTensor, in particular, it:
- allows DTensor be passed into a compiled function, and allow fakify
DTensor during dynamo tracing by turning the inner local tensor to meta
tensor.
- We use `allow_in_graph` to include `DTensor` and `DTensor.from_local` to be represented as `TorchVariable`
- The dtensor created becomes a normal `TensorVariable` and it would insert any tensor operations to the output graph just like torch.Tensor
- note that dtensor have a new instance method `redistribute` compare to plain tensor, and we currently special handle it in `TensorVariable`
`from_local` and `redistribute` both accepts some non-trival metadata as arguments (i.e. DeviceMesh, Placement) which fx.Graph does not support. In order to let these two APIs appear in the dynamo captured graph, we encoded the metadata into a new_function (like `functools.partial`) and the new function only accepts prim args (i.e. tensor), then we put `call_function` with this new_function to the graph. This is suggested by @ezyang. The underlying rationale here is that the metadata will not change across the graph invocations so it's safe to encode them.
Captured graph:
```
def forward(self, L_x_ : torch.Tensor):
l_x_ = L_x_
# File: /scratch/wanchaol/work/pytorch/test/distributed/_tensor/test_dtensor.py:685, code: dt = DTensor.from_local(x, mesh, [Shard(0)], run_check=False)
prim_from_local = torch__dynamo_variables_torch_prim_from_local(l_x_, run_check = False); l_x_ = None
# File: /scratch/wanchaol/work/pytorch/test/distributed/_tensor/test_dtensor.py:686, code: return dt.redistribute(mesh, [Replicate()]).to_local() + 2
prim_redistribute = torch__dynamo_variables_tensor_prim_redistribute(prim_from_local); prim_from_local = None
to_local = prim_redistribute.to_local(); prim_redistribute = None
add = to_local + 2; to_local = None
return (add,)
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/103146
Approved by: https://github.com/voznesenskym
The main concept behind this refactor is this: if we know that a size/stride/etc is constant, do NOT trace it into the graph, EXCEPT for any preexisting special cases that applied for static shapes. The refactor unfolds like this:
1. Delete the `dynamic_shapes` branches in torch/_dynamo/variables/builder.py which accept int/float/bool outputs. This is over-aggressive and we don't want to allow this (because if the operator returns a constant, we shouldn't have called wrap_fx_proxy in the first place.) This causes a bunch of failures because we are blindly feeding the result of size() call to wrap_fx_proxy when dynamic shapes is enabled.
2. Modify TensorVariable.call_method in torch/_dynamo/variables/tensor.py to avoid sending constant ints to wrap_fx_proxy. After normal specialization (which should be deleted, see https://github.com/pytorch/pytorch/pull/103434) we consult the fake tensor to see if the values in question have free variables or not. If they don't we short circuit tracing into graph. We only trace into graph if the operation in question is truly symbolic. Note that there is a near miss here: it's OK to trace x.size() call entirely into the graph, even if it doesn't have all dynamic shapes, because operator.getitem with int output is special cased in builder.py. This is a preexisting special case and I don't try to get rid of it.
3. It turns out that the change here also breaks torch_np compatibility layer. So I completely rewrite getattr handling in torch/_dynamo/variables/tensor.py to follow the same pattern (only trace into graph if truly dynamic).
There's some minor housekeeping in torch/fx/experimental/symbolic_shapes.py and some test files.
Signed-off-by: Edward Z. Yang <ezyang@meta.com>
Pull Request resolved: https://github.com/pytorch/pytorch/pull/103438
Approved by: https://github.com/larryliu0820
This PR adds universal support for ndarray methods. After #100839 each `NumpyNdarrayVariable` should wrap a `torch.Tensor`. This PR adds a `numpy_method_wrapper` which converts the `torch.Tensor` to `torch_np.ndarray` and then call the numpy ndarray method. Then we also try to return a `torch.Tensor` (return as-is if the value is not ndarray-like)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/97537
Approved by: https://github.com/ezyang
Previously I accidentally thought setitem takes each argument as a
list. But if you write x[:, b] that actually is passed in as a tuple.
Try harder.
Signed-off-by: Edward Z. Yang <ezyang@meta.com>
Pull Request resolved: https://github.com/pytorch/pytorch/pull/103052
Approved by: https://github.com/desertfire
Issue: #93684
In previous PRs #95849#99560 we redirect `numpy.*`, `<tensor>.numpy()` calls to `torch_np.*` methods and attributes, by creating `NumpyNdarrayVariable` for those calls.
We need to handle `NumpyNdarrayVariable` when graph break happens.
This PR did 2 things:
1. In `codegen.py` we made sure we can reconstruct the value wrapped by `NumpyNdarrayVariable`, to be `torch_np.ndarray` in the stack whenerver we recompiles the subgraph.
2. In `builder.py` we can wrap the value to be `NumpyNdarrayVariable` and save it as graph input.
-----
Starting from commit 6:
## A new design for supporting numpy in dynamo
In short the core concept doesn't change: we still convert `numpy` API calls to `torch_np` API calls. However, instead of wrapping a `torch_np.ndarray` in `NumpyNdarrayVariable`, the new design wraps a `torch.Tensor`.
The reason for doing this change is because we need to keep `torch.Tensor` everywhere in the captured graph, so that it works well with the backend of dynamo. See discussions in https://github.com/Quansight-Labs/numpy_pytorch_interop/issues/142 for details.
### Flow
This is an example showing how do we think about dynamo working on a simple function:
```python
def f(x: torch.Tensor, y: torch.Tensor):
a, b = x.numpy(), y.numpy()
c = np.add(x, y)
return torch.from_numpy(c)
```
```
+------------+ +------------+
torch.Tensor | |numpy.ndarray| |
-------------- .numpy() --------------| |
| | | | +------------------+
+------------+ | numpy.add |numpy.ndarray| |torch.Tensor
+------------+ | --------------| torch.from_numpy --------------
torch.Tensor | |numpy.ndarray| | | |
-------------- .numpy() --------------| | +------------------+
| | | |
+------------+ +------------+
+------------+ +----------------+
torch.Tensor | |torch.Tensor | |
-------------- .detach() --------------| |
| | | | +----------------+ +------------+
+------------+ | |torch_np.ndarray| |torch.Tensor| |torch.Tensor
| torch_np.add -----------------| util.to_tensor -------------| .detach() --------------
+------------+ | | | | | |
torch.Tensor | |torch.Tensor | | +----------------+ +------------+
-------------- .detach() --------------| |
| | | |
+------------+ | +----------------+ |
| wrapper on torch_np.add |
+--------------------------------------------------------+
```
### Approach
`torch_np` APIs can take both `torch_np.ndarray` as well as `torch.Tensor`. What we need to do is to have a wrapper for these APIs to convert the return value back to `torch.Tensor`. This way only the wrapper is showing up in the captured graph, with `torch.Tensor`s as input and `torch.Tensor` as output.
If we have a graph break or we've traced to the end of the program, we need to inspect all the `NumpyNdarrayVariable` in the stack and convert them back to `numpy.ndarray`, to make sure the compiled version is still behaving the same as the eager version.
### Examples
Here's an example of the graph generated:
```python
def fn(x: np.ndarray, y: np.ndarray):
a = x.real
b = y.real
torch._dynamo.graph_break()
return np.add(a, 1), np.add(b, 1)
```
Graph generated:
```
[2023-05-16 10:31:48,737] torch._dynamo.output_graph.__graph: [DEBUG] TRACED GRAPH
__compiled_fn_0 <eval_with_key>.0 opcode name target args kwargs
------------- -------------- ---------------------------------------------------------- ---------------------- --------
placeholder l_x_ L_x_ () {}
placeholder l_y_ L_y_ () {}
call_function from_numpy <built-in method from_numpy of type object at 0x12b1fdc80> (l_x_,) {}
call_function from_numpy_1 <built-in method from_numpy of type object at 0x12b1fdc80> (l_y_,) {}
call_function attr_wrapper <function attr_wrapper at 0x12e8693a0> (from_numpy, 'real') {}
call_function attr_wrapper_1 <function attr_wrapper at 0x12e8693a0> (from_numpy_1, 'real') {}
output output output ((),) {}
[2023-05-16 10:31:48,908] torch._dynamo.output_graph.__graph: [DEBUG] TRACED GRAPH
__compiled_fn_2 <eval_with_key>.1 opcode name target args kwargs
------------- ------------- ---------------------------------------------------------- ------------------------------- --------
placeholder l_a_ L_a_ () {}
placeholder l_b_ L_b_ () {}
call_function from_numpy <built-in method from_numpy of type object at 0x12b1fdc80> (l_a_,) {}
call_function from_numpy_1 <built-in method from_numpy of type object at 0x12b1fdc80> (l_b_,) {}
call_function wrapped_add <Wrapped function <original add>> (from_numpy, 1) {}
call_function wrapped_add_1 <Wrapped function <original add>> (from_numpy_1, 1) {}
output output output ((wrapped_add, wrapped_add_1),) {}
```
### Changes
* `codegen.py`: reconstruct `numpy.ndarray` from `NumpyNdarrayVariable` by adding bytecode to call `utils.to_numpy_helper()`.
* `output_graph.py`: getting rid of legacy code that does exactly what `codegen.py` does, which only handling return case but not graph break case.
* `utils.py`: added helpers to convert `numpy.ndarray` to `torch.Tensor` and vice versa. Also adding a wrapper class that takes in a function. In `__call__` it calls the function and converts its out to `torch.Tensor` (or a list of it).
* `builder.py`: add method to wrap `numpy.ndarray` graph inputs into `NumpyNdarrayVariable`, by calling `torch.numpy` in the proxy.
* `misc.py`: `numpy` API calls goes into `NumpyVariable` and we find the function with the same name in `torch_np` module, then wrap it with the wrapper defined in `utils.py`.
* `tensor.py`, `torch.py`: proxy `tensor.numpy()` to be `torch.detach()` but wrap it with `NumpyNdarrayVariable`. Similarly, `torch.from_numpy()` -> `torch.detach()` but wrap it with `TensorVariable`. In `NumpyNdarrayVariable`, do the similar `torch_np.ndarray` to `torch.Tensor` wrapping for attributes.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/100839
Approved by: https://github.com/ezyang
This PR adds support for tracing autograd.Function with grad.
A few important bullet points outlining our approach:
1) Our goal is to verify soundness in order to add a call_function to the autograd.Function's `apply` to the graph.
2) We achieve (1) by either verifying soundness or rejecting soundness, by ensuring that both forward and backward of the autograd.Function are sound.
3) For the forward, if we verify soundness, we install its guards into the graph.
4) For the backward, if we verify soundness, we throw it out. However, backwards soundness verification is more onerous, and has a config driven set of banned attrs and methods for tensors.
1-4 above are achieved by turning the forward and backward into UserDefinedFunctionVariables, and inlining through them, relying on dynamo's soundness detection. If we graph break in these, we raise and treat them as unsound. As noted above, backwards is stricter yet.
For the tracing, the safety comes from dynamo's HigherOrderOperator system. That system ensures that not only do we trace soundly, but that no new variables are lifted into inputs during the tracing, and that the forward and backwards are entirely self contained.
Whenever we reject a function as unsound, we restore back, as usual.
Due to some limitations in the lifting logic, we have an escape hatch we implemented for tensors that are known in forward, but cross into backwards through save_tensors (save) /saved_tensors (load). We escape hatch here to avoid having the known saved tensors coming from forward end up being accidentally treated as lifted variables (and rejected). This is sound, but a little hacky feeling.
Additionally, due to some limitations in fx node removal, combined with how we produce subgraphs for the traces installed from HigherOrderOperators, we had to improve our node removal logic. In the event of a restore, we remove the old nodes from the graph, as usual in dynamo. However, because the references to these nodes may exist in subgraphs, we traverse any nodes users and remove them first if and only if they are in another graph. This is always sound, because removal should only be downstream of restoration at this point.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/99483
Approved by: https://github.com/zou3519
