Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/50393
Exponential Moving Average
Usage:
add ema_options in adagrad optimizer. For details, plz refer to the test workflow setting.
if ema_end == -1, it means ema will never end.
Test Plan:
buck test caffe2/caffe2/fb/optimizers:ema_op_optimizer_test
buck test caffe2/caffe2/fb/optimizers:ema_op_test
f240459719
Differential Revision: D25416056
fbshipit-source-id: a25e676a364969e3be2bc47750011c812fc3a62f
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/49486
Remove code for Python 3.5 and lower.
There's more that can be removed/modernised, but sticking mainly to redundant version checks here, to keep the diff/PR smaller.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/46579
Reviewed By: zou3519
Differential Revision: D24453571
Pulled By: ezyang
fbshipit-source-id: c2cfcf05d6c5f65df64d89c331692c9aec09248e
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/49591
A bunch of these tests are marked flaky, and have been since time immemorial. (Read: as far back as Buck will build.) However closer inspection reveals that they fail if and only if run on a GPU worker. What seems to be going on is that there are more jobs than GPUs, so the contention causes waits which registers as timeouts on the test.
This diff is kind of hacky, but it basically just drops deadlines if a GPU is present. Because Caffe2 is going away I'm not too terribly concerned about a beautiful solution, but we may as well keep some test coverage if it's easy.
CC Sebastian, Ilia, Min, and Hongzheng who also have tasks for what seems to be the same flakiness.
Test Plan: Turn the tests back on and see if they fall over. (The failure repros reliably on an OnDemand GPU and is fixed by this change, so it's not really just a hail Mary.)
Reviewed By: ngimel
Differential Revision: D25632981
fbshipit-source-id: 43dcce416fea916ba91f891e9e5b59b2c11cca1a
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/49402
In cases of NCCLAllReduce operations there could be non-trivial overhead for
launching cooperative kernels (especially in case of async execution of
different parts of the model). This diff is reviving this operator to make it
possible to fuse multiple operations into a single kernel.
Test Plan:
Unit-test.
Used in a later diff.
Reviewed By: xianjiec
Differential Revision: D25531206
fbshipit-source-id: 64b1c161233a726f9e2868f1059316e42a8ea1fc
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/49322
In some cases async execution might loose dependencies (Alias like ops) or produce suboptimal scheduling when there is an option which parts to schedule first. Example of the later behavior can happen in ModelParallel training where copy can get lower priority compared to the rest of the execution on the given GPU, which will caused other GPUs to starve.
This operator allows to address these issues by introducing extra explicit dependencies between ops.
Test Plan:
Unit-test/
E2E testing in the future diffs.
Reviewed By: xianjiec
Differential Revision: D24933471
fbshipit-source-id: 1668994c7856d73926cde022378a99e1e8db3567
Summary:
+ Add ArgMin support to Caffe2 to PyTorch converter
+ Using hypothesis to parameterize different conditions for test
Test Plan: buck test //caffe2/torch/fb/model_transform/c2_convert:c2_pt_converter_test
Reviewed By: houseroad
Differential Revision: D25016203
fbshipit-source-id: 94489fcf1ed3183ec96f9796a5b4fb348fbde5bc
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/48240
Adds the support for converting the SparseLengthsSum4BitRowwiseSparse operator from caffe2 to pytorch as a part of c2_pt_converter
Test Plan:
Added a unit tested
buck test //caffe2/torch/fb/model_transform/c2_convert:c2_pt_converter_test
Tests Passed :
https://our.intern.facebook.com/intern/testinfra/testrun/2251799856412296
Reviewed By: houseroad
Differential Revision: D25067833
fbshipit-source-id: 45cbc331ca35bee27e083714e65a1e87a2a2d2e0
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/48340
This changes the context managed classes from using a decorator to define them to using inheritance. Inheritance allows the python static type checking to work correctly.
```
context.define_context()
class Bar(object): ...
context.define_context(allow_default=True)
class Foo(object): ...
```
becomes
```
class Foo(context.Managed): ...
class Bar(context.DefaultManaged): ...
```
Behavior differences:
* arg_name has been removed since it's not used anywhere
* classes need to call `super()` in `__enter__/__exit__` methods if they override (none do)
This also defines a context.pyi file to add types for python3. python2 support should not be affected
Test Plan:
ci
buck test //caffe2/caffe2/python:context_test //caffe2/caffe2/python:checkpoint_test
Reviewed By: dongyuzheng
Differential Revision: D25133469
fbshipit-source-id: 16368bf723eeb6ce3308d6827f5ac5e955b4e29a
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/48407
T79817692: Fused8BitRowwiseQuantizedToFloat operator support for c2_pt_converter.
Also refactored some repeated code from the existing test functions. (Initial commit only has refactoring.)
Test Plan: buck test //caffe2/torch/fb/model_transform/c2_convert:c2_pt_converter_test
Reviewed By: bugra
Differential Revision: D25069936
fbshipit-source-id: 72f6a845a1b4639b9542c6b230c8cd74b06bc5a0
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/48404
On bento this is printing a lot of msgs like (see N408483 if you're an internal user)
```
W1123 120952.322 schema.py:811] Scalar should be considered immutable. Only call Scalar.set() on newly created Scalar with unsafe=True. This will become an error soon.
```
And it's ignoring the log level I set at global level. Removing this line unless this is super important.
Test Plan: build a local dper package and verify
Differential Revision: D25163808
fbshipit-source-id: 338d01c82b4e67269328bbeafc088987c4cbac75
Summary: is_external_input doesn't check if the lookup tables are valid. Calling .Proto() should invalidate all lookup tables and have them rebuilt on call to any methods depending on them. This adds this check to is_external_input.
Test Plan: internal unit tests
Reviewed By: dzhulgakov, esqu1
Differential Revision: D25100464
fbshipit-source-id: d792dec7e5aa9ffeafda88350e05cb757f4c4831
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/47023
DeviceType pretty clearly only needs 1 byte. DeviceIndex only needs 1 byte given that machines don't have anywhere near 255 GPUs in them as far as I know.
ghstack-source-id: 116901430
Test Plan: Existing tests, added assertion to catch if my assumption about DeviceIndex is incorrect
Reviewed By: dzhulgakov
Differential Revision: D24605460
fbshipit-source-id: 7c9a89027fcf8eebd623b7cdbf6302162c981cd2
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/47768
This stores the next ID for a given NextName(prefix, output_id) so repeated calls to NextName are significantly faster. This accounts for ~65% of time spent for large models.
Test Plan:
buck test //caffe2/caffe2/python/...
will launch canary job before landing to ensure no regressions + confirm speedup
Reviewed By: dzhulgakov
Differential Revision: D24876961
fbshipit-source-id: 668d73060d800513bc72d7cd405a47d15c4acc34
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/48021
Extending operator schema check for simple memonger to dag memonger as well. As part of this a fix is being made to handle inplace ops (having at least one output name same as input blob). Earlier all the output blobs from ops were being treated as shareable but it failed assertion of external input blobs with the same name not allowed to share.
Test Plan: Added corresponding unit tests
Reviewed By: hlu1
Differential Revision: D24968862
fbshipit-source-id: b6679a388a82b0d68f65ade64b85560354aaa3ef
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/47718
Distributed Inference splits a predict net into multiple parts, part0 being the main part which contains ops to make remote calls to other parts. part0 predict net may contain AsyncIf ops to optimize rpc call usage. AsyncIf ops have internal nets which may refer to memongered blobs. This change handles AsyncIf ops to update internal nets to refer to memongered blobs.
As part of this change, I am also updating dag memonger traversal to always start from root op, i.e. ops with 0 in degree. Earlier logic will start traversing ops based on input head blobs and if one of the head inputs is getting used in a non-root op which gets visited before its parent, the traversal will throwing assertion error here: https://fburl.com/diffusion/ob110s9z . Almost for all the distributed inference part0 nets, it was throwing this assertion error.
Test Plan: Added corresponding tests in memonger_test.py . Could not find unit tests in c++ version of memonger.
Reviewed By: hlu1
Differential Revision: D24872010
fbshipit-source-id: 1dc99b2fb52b2bc692fa4fc0aff6b7e4c5e4f5b0
Summary: Added the MatMul operator for caffe2
Test Plan: buck test //caffe2/torch/fb/model_transform/c2_convert:c2_pt_converter_test
Reviewed By: bugra
Differential Revision: D24920937
fbshipit-source-id: 7ba09ba0439cb9bd15d6a41fd8ff1a86d8d11437
Summary: To support min/max/mean/std, SummarizeOp need to skip size checking (similar to the LpNorm error mentioned above) and accept multiple types
Test Plan:
unit test:
`buck test //caffe2/caffe2/fb/tensorboard/tests:tensorboard_accumulate_histogram_op_test`
https://our.intern.facebook.com/intern/testinfra/testrun/1407375057859572
`buck test //caffe2/caffe2/fb/tensorboard/tests:tensorboard_accumulate_histogram_op_test --stress-runs 1000`
https://our.intern.facebook.com/intern/testinfra/testrun/2533274832166362
Reviewed By: cryptopic
Differential Revision: D24605507
fbshipit-source-id: fa08372d7c9970083c38abd432d4c86e84fb10e0
Summary:
Distributed Inference splits a predict net into multiple parts, part0 being the main part which contains ops to make remote calls to other parts. part0 predict net may contain AsyncIf ops to optimize rpc call usage. AsyncIf ops have internal nets which may refer to memongered blobs. This change handles AsyncIf ops to update internal nets to refer to memongered blobs. Here is one reference part0 predict net with AsyncIf ops: https://www.internalfb.com/intern/paste/P145812115/
As part of this change, I am also updating dag memonger traversal to always start from root op, i.e. ops with 0 in degree. Earlier logic will start traversing ops based on input head blobs and if one of the head inputs is getting used in a non-root op which gets visited before its parent, the traversal will throwing assertion error here: https://fburl.com/diffusion/ob110s9z . Almost for all the distributed inference part0 nets, it was throwing this assertion error.
Reviewed By: hlu1
Differential Revision: D24346771
fbshipit-source-id: ad2dd2e63f3e822ad172682f6d63f8474492255d
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/47541
The profiler has guided us to `schema.py`. Since these `Field`s are used everywhere and in huge quantities, we can easily make some optimizations system wide by adding `__slots__`.
From StackOverflow, benefits include:
* faster attribute access.
* space savings in memory.
Read more: https://stackoverflow.com/a/28059785/
Reviewed By: dzhulgakov
Differential Revision: D24771078
fbshipit-source-id: 13f6064d367440069767131a433c820eabfe931b
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/47542
The previous way of doing `Field.__init__(self, [])` is just wrong. Switching to Python2 compatible way: `super(ObjectName, self).__init__(...)`
Reviewed By: dzhulgakov
Differential Revision: D24771077
fbshipit-source-id: d6798c72090c0264b6c583602cae441a1b14587c
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/47530
`Net.AddExternalInput` should raise if there are duplicate names. The previous code would only raise if the addition of duplicates was in separate calls, but not if it was in the same call.
Test Plan:
Added two new regression tests
```
✓ Pass: caffe2/caffe2/python:core_test - testSetInputRecordWithBlobs (caffe2.caffe2.python.core_test.TestExternalInputs) (9.622)
✓ Pass: caffe2/caffe2/python:core_test - testAddExternalInputShouldRaiseIfDuplicate (caffe2.caffe2.python.core_test.TestExternalInputs) (9.639)
✓ Pass: caffe2/caffe2/python:core_test - testSetInputRecordWithoutBlobs (caffe2.caffe2.python.core_test.TestExternalInputs) (9.883)
✓ Pass: caffe2/caffe2/python:core_test - testAddExternalInputShouldRaiseIfDuplicateInSameCall (caffe2.caffe2.python.core_test.TestExternalInputs) (10.153)
```
Test trained 2 models. No issues
f230755456
f230754926
Reviewed By: dzhulgakov
Differential Revision: D24763586
fbshipit-source-id: c87088441d76f7198f8b07508b2607aec13521ed
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/47512
I deleted the last line of `__init__` -- `self._field_offsets.append(offset)` -- and the unittests didn't fail.
So this diff is to improve test coverage.
Test Plan:
```
✓ Pass: caffe2/caffe2/python:schema_test - testInitShouldSetEmptyParent (caffe2.caffe2.python.schema_test.TestField) (8.225)
✓ Pass: caffe2/caffe2/python:schema_test - testInitShouldSetFieldOffsetsIfNoChildren (caffe2.caffe2.python.schema_test.TestField) (8.339)
✓ Pass: caffe2/caffe2/python:schema_test - testInitShouldSetFieldOffsets (caffe2.caffe2.python.schema_test.TestField) (8.381)
```
Reviewed By: dzhulgakov
Differential Revision: D24767188
fbshipit-source-id: b6ce8cc96ecc61768b55360e0238f7317a2f18ea
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/47475
This improves the core.Net cloning/init performance by quite a bit. It makes set_input_record run in linear time instead of O(n) by checking the external_input map instead of regenerating the external inputs each time and then iterating over it.
Test Plan: unit tests + canary runs
Reviewed By: dzhulgakov
Differential Revision: D24765346
fbshipit-source-id: 92d9f6dec158512bd50513b78675174686f0f411
Summary:
Add `last_n_window_collector` as C2 supports and PyTorch currently does not have this operator: https://www.internalfb.com/intern/diffusion/FBS/browsefile/master/fbcode/caffe2/caffe2/operators/last_n_window_collector.cc?lines=139
## Problem that we are solving
This operator works on multiple pieces of data and collects last `n` element that has been seen.
If you have the following pieces of data that has been passed around:
```
[1, 2, 3, 4]
[5, 6, 7]
[8, 9, 10, 11]
```
for 3 times and the number of collector is given to be 6. The expected result is:
```
[6, 7, 8, 9, 10, 11]
```
What this means is that, almost like we need a FIFO(First in First Out) mechanism where as we are passing this data through the collector, we will be pushing some other data at the end.
In this particular example, in the first pass(the data is `[1, 2, 3, 4]`) , we hold `[1, 2, 3, 4]` in the queue as our queue size is 6.
In the second pass(the data is `[5, 6, 7]`), we hold `[2, 3, 4, 5, 6, 7]` in the queue and since 1 is inserted the last, it will drop due to the size limitation of the queue.
In the third pass(the data is `[8, 9, 10, 11]`), we hold `[6, 7, 8, 9, 10, 11]` in the queue and `2,3,4,5` are dropped due the the size of the queue.
For multidimension case, when we have the following data:
```
[[1, 2], [2, 3], [3, 4], [4, 5]]
[[5, 6], [6, 7], [7, 8]]
[[8, 9], [9, 10], [10, 11], [11, 12]]
```
and our queue size is 6.
In the first pass, we will have ` [[1, 2], [2, 3], [3, 4], [4, 5]]`
In the second pass, we will have `[2, 3], [3, 4], [4, 5]] [[5, 6], [6, 7], [7, 8]]`
In the third pass, we will have `[6, 7], [7, 8]] [[8, 9], [9, 10], [10, 11], [11, 12]]`
### The implementation
I am using FIFO queue in Python which is in the collections library. This accepts `maxlen` argument which can be used to set the size of the queue.
I am using last n indices of the tensor through list indices and in this operator, I am not doing copy.
In the test plan, I have both single dimension tensors as well as multi-dimension tensors.
### Benchmark
I used various different configurations and added a benchmark test. PyTorch implementation is much master than Caffe2 implementation:
#### CPU Benchmark
```
torch_response.median
0.00019254473969340324
caffe_response.median
0.00030233583599794657
```
#### GPU Benchmark
```
torch_response.mean
0.000081007429903838786
caffe_response.mean
0.00010279081099724863
```
Test Plan:
### For CPU:
```
buck test //caffe2/torch/fb/sparsenn:test
```
### For GPU:
- Used an on-demand machine and did the following commands:
```
jf get D24435544
buck test mode/opt //caffe2/torch/fb/sparsenn:test
```
https://www.internalfb.com/intern/testinfra/testconsole/testrun/4222124688138052/
Reviewed By: dzhulgakov, radkris-git
Differential Revision: D24435544
fbshipit-source-id: 8193b4746b20f2a4920fd4d41271341045cdcee1
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/46590
This operator is very similar to LengthsToRanges but doesn't pack the offsets next to the original lengths.
Reviewed By: yf225
Differential Revision: D24419746
fbshipit-source-id: aa8b014588bb22eced324853c545f8684086c4e4
Summary: I was reading/looking into how LocalSession works and realized that the workspace type being passed around was the bound function on TaskGroup instead of the actual type. This meant that all workspaces for localsession would always be global, because they'd never match the private workspace type.
Test Plan: <not sure, could use some suggestions>
Reviewed By: cryptopic
Differential Revision: D24458428
fbshipit-source-id: 0f87874babe9c1ddff25b5363b443f9ca37e03c1
Summary:
We've been seeing a lot of Hypothesis timeouts and from profiling a few of the failing tests one of the contributing factors is really slow grad checker. In short, it launches the whole op for each of the input elements so the overall complexity is O(numel^2) at least.
This applies a very unscientific hack to just run grad check on the first and last few elements. It's not ideal, but it's better than flaky tests. One can still explicitly opt in with the env var.
Reviewed By: malfet
Differential Revision: D23336220
fbshipit-source-id: f04d8d43c6aa1590c2f3e72fc7ccc6aa674e49d2
Summary: Similar to If operator, AsyncIf also contains nets in args. It needs the same handling.
Test Plan:
New unit test test_control_op_remap
`buck test caffe2/caffe2/python:core_test`
Also it worked end to end in prototype of dist bulk eval workflow f226680903
Reviewed By: yyetim
Differential Revision: D24451775
fbshipit-source-id: 50594e2ab9bb457329ed8da7b035f7409461b5f6
Summary:
Follow-up of https://github.com/pytorch/pytorch/issues/46461 with a similar goal
Makes them more readable and possibly faster. Care has to be taken because `map` applies the function immediately while `(x for x in xs)` is a generator expression which gets evaluated later. This is a benefit in some cases where it is not required to actually create the list of values in memory (e.g. when passing to `tuple` or `extend` or `join`)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/46462
Reviewed By: zou3519
Differential Revision: D24422343
Pulled By: ezyang
fbshipit-source-id: 252e33499c92ac0b15238f2df32681dbbda2b237
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/46457
Wanted to see if using CopyMatrix specialized for float that uses mkl_somatcopy can be faster but it wasn't. Still want to check in benchmark that can be used later.
Test Plan: .
Reviewed By: dskhudia
Differential Revision: D24345901
fbshipit-source-id: d3e68dbb560e3138fda11c55789cd41bc0715c6d
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/45551
The FP16 version of SparseNormalize op in Caffe2 is missing. This Diff adds FP16 support to unblock MC process of adding FP16 to Dper3.
Check https://fb.quip.com/L0T2AXGwUY3n#EReACAeifk3 .
One question is whether the pure FP16 Sparse Normalized op will affect the accuracy? Maybe we should do it in FP32 domain.
ghstack-source-id: 114184398
Test Plan:
```
buck run mode/opt //caffe2/caffe2/python/operator_test:sparse_normalize_test
```
```
buck run mode/opt -c python.package_style=inplace mode/no-gpu //caffe2/caffe2/python/benchmarks:sparse_normalize_benchmark -- --fp16
```
Reviewed By: jspark1105
Differential Revision: D24005618
fbshipit-source-id: 8b918ec4063fdaafa444779b95206ba2b7b38537
Summary: This diff adds a string equality checking operator.
Test Plan: Unit tests
Differential Revision: D24042344
fbshipit-source-id: c8997c6130e3438f2ae95dae69f76978e2e95527
Summary: `__repr__` calling self.tasks() ends up marking the instance as "used", which doesn't seem appropriate. I was debugging a value being passed around and then ran into `Cannot add Task to an already used TaskGroup.` because the value had been logged once.
Test Plan:
Added a unit test -- didn't see a clean public method to test it, but I'm happy to add one if that makes sense.
Will wait for sandcastle to trigger everything else; I'm not at all familiar with this code so any other recommendations would be great!
Reviewed By: cryptopic
Differential Revision: D23541198
fbshipit-source-id: 5d1ec674a1ddaedf113140133b90e0da6afa7270
Summary: Currently GetSingleArgument is overflowing since it's expecting an int instead of an int64 when using a 1cycle (hill policy) annealing schedule
Test Plan:
unittest
buck test caffe2/caffe2/python/operator_test:learning_rate_op_test
Differential Revision: D23938169
fbshipit-source-id: 20d65df800d7a0f1dd9520705af31f63ae716463
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/45315
Pull Request resolved: https://github.com/pytorch/pytorch/pull/45314
in D23858329 (721cfbf842), we put PriorCorrectionCalibrationPrediction unit test in OSS file which causes test failure issue in public trunk.
this diff moves it to FB only test file.
Test Plan:
```
buck test //caffe2/caffe2/python/operator_test:torch_integration_test -- test_gather_ranges_to_dense_op
buck test //caffe2/caffe2/fb/python/operator_test:torch_integration_test -- test_prior_correct_calibration_prediction_op
```
all pass.
Reviewed By: houseroad
Differential Revision: D23899012
fbshipit-source-id: 1ed97d8702e2765991e6caf5695d4c49353dae82
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/45178
## Motivation
* To be able to make C2 ops cancellable so we can safely exit.
* Some C2 operators are now blocking thus being non-cancellable. If an error
occurs we need to be able to safely stop all net execution so we can throw
the exception to the caller.
## Summary
* Adds a hypothesis test for queue ops cancellation.
Test Plan:
## Unit test added to verify that queue ops propagate errors
```
buck test caffe2/caffe2/python:hypothesis_test
buck test caffe2/caffe2/python:hypothesis_test -- test_safe_dequeue_blob__raises_exception_when_hang --stress-runs 1000
```
```
Summary
Pass: 1000
ListingSuccess: 1
```
Reviewed By: d4l3k
Differential Revision: D23847576
fbshipit-source-id: 2fc351e1ee13ea8b32d976216d2d01dfb6fcc1ad
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/45231
There are two operators:
`PriorCorrectionCalibrationPrediction` and `GatherRangesToDense` is not supported in PT which makes GLOW cannot work.
To unblock, we first try to use C2->PT conversion. In the long-term, we need to implement PT custom ops.
This diff does this conversion to unblock current project.
Test Plan:
Run unit test. the Test input is from current DPER example.
All pass.
```buck test //caffe2/caffe2/python/operator_test:torch_integration_test -- test_prior_correct_calibration_prediction_op --print-passing-details
> c2 reference output
> [0.14285715 0.27272728 0.39130434 0.5 ]
> PT converted output
> tensor([0.1429, 0.2727, 0.3913, 0.5000])
buck test //caffe2/caffe2/python/operator_test:torch_integration_test -- test_gather_ranges_to_dense_op --print-passing-details
c2 reference output
> [array([[6, 5, 4, 3], [0, 0, 0, 0]], dtype=int64)]
> PT converted output
> [tensor([[6, 5, 4, 3], [0, 0, 0, 0]])]
```
Reviewed By: allwu, qizzzh
Differential Revision: D23858329
fbshipit-source-id: ed37118ca7f09e1cd0ad1fdec3d37f66dce60dd9
Summary:
There is a module called `2to3` which you can target for future specifically to remove these, the directory of `caffe2` has the most redundant imports:
```2to3 -f future -w caffe2```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/45033
Reviewed By: seemethere
Differential Revision: D23808648
Pulled By: bugra
fbshipit-source-id: 38971900f0fe43ab44a9168e57f2307580d36a38
Summary:
## Motivation
* To be able to make C2 ops cancellable so we can safely exit.
* Some C2 operators are now blocking thus being non-cancellable. If an error
occurs we need to be able to safely stop all net execution so we can throw
the exception to the caller.
* When an error occurs in a net or it got cancelled, running ops will have the
`Cancel` method called.
* This diff adds `Cancel` method to the `SafeEnqueueBlobsOp`
and `SafeDequeueBlobsOp` to have the call queue->close() to force all the
blocking ops to return.
* Adds unit test that verified the error propagation.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/44495
Test Plan:
## Unit Test added to verify that queue ops propagate errors
```
buck test caffe2/caffe2/python:hypothesis_test
```
Reviewed By: dzhulgakov
Differential Revision: D23236088
Pulled By: dahsh
fbshipit-source-id: daa90d9ee32483fb51195e269a52cf5987bb0a5a
Summary:
Make `gcs_cuda_only` and `gcs_gpu_only` return empty device lists if CUDA/GPU(CUDA or RocM) not available
Pull Request resolved: https://github.com/pytorch/pytorch/pull/44578
Reviewed By: walterddr
Differential Revision: D23664227
Pulled By: malfet
fbshipit-source-id: 176b5d964c0b02b8379777cd9a38698c11818690
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/44540
Support output type to be fp16 for UniformFill
Reviewed By: jianyuh
Differential Revision: D23558030
fbshipit-source-id: 53a5b2c92cfe78cd11f55e6ee498e1bd682fe4a1
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/44089
Add support of fp16 as input type in SparseLengthSum/Mean caffe2 operator
Reviewed By: xianjiec
Differential Revision: D23436877
fbshipit-source-id: 02fbef2fde17d4b0abea9ca5d17a36aa989f98a0
Summary:
Expose the interface of `nesterov` of SGD Optimizer from caffe2 to dper.
dper sgd optimizer (https://fburl.com/diffusion/chpobg0h) has referred to NAG sgdoptimizer in caffe2: https://fburl.com/diffusion/uat2lnan. So just need to add the parameter 'nesterov' in dper sgd optimizer.
Analysis of run resutls: N345540.
- train_ne increases as momentum (m) decreases.
- for m=0.95, 0.9: eval_ne is lower with NAG than production (no NAG, m = 0.95).
- for m=0.99: eval_ne with or without NAG is higher than production. It indicates larger variance in validation and overfit in training (lower train_ne).
Test Plan:
1. unit tests:
`buck test caffe2/caffe2/fb/dper/layer_models/tests/split_1:sparse_nn_test -- test_sgd_without_nesterov`
`buck test caffe2/caffe2/fb/dper/layer_models/tests/split_1:sparse_nn_test -- test_sgd_with_nesterov`
.
1. build dper front end package: `flow-cli canary ads.dper3.workflows.sparse_nn.train --mode opt --entitlement ads_global --run-as-secure-group team_ads_ml_ranking`. The build result (refreshed) is here https://www.internalfb.com/intern/buck/build/2a368b55-d94b-45c1-8617-2753fbce994b. Flow package version is ads_dper3.canary:856b545cc6b249c0bd328f845adeb0d2.
.
2. To build dper back end package: `flow-cli canary dper.workflows.dper3.train --mode opt --entitlement ads_global --run-as-secure-group team_ads_ml_ranking`. The build result (refreshed) is here: https://www.internalfb.com/intern/buck/build/70fa91cd-bf6e-4a08-8a4d-41e41a77fb52. Flow package version is aml.dper2.canary:84123a34be914dfe86b1ffd9925869de.
.
3. Compare prod with NAG-enabled runs:
a) refreshed prod run (m=0.95): f213877098
NAG enabled run (m=0.95): f213887113
.
b) prod run (m=0.9): f214065288
NAG enabled run (m=0.9): f214066319
.
c) prod run (m=0.99): f214065804
NAG enabled run (m=0.99): f214066725
.
d) change date type of nestrov to `bool` and launched a validation run
NAG enabled (m=0.95): f214500597
Reviewed By: ustctf
Differential Revision: D23152229
fbshipit-source-id: 61703ef6b4e72277f4c73171640fb8afc6d31f3c
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/44043
To invoke `cancel` from the net instance in Python, we expose it through pybind state.
Reviewed By: dzhulgakov
Differential Revision: D23249660
fbshipit-source-id: 45a1e9062dca811746fcf2e5e42199da8f76bb54
Summary: Exports the operator to PyTorch, to be made into a low-level module.
Test Plan:
```
buck test //caffe2/caffe2/python/operator_test:torch_integration_test -- test_learning_rate
```
Reviewed By: yf225
Differential Revision: D23545582
fbshipit-source-id: 6b6d9aa6a47b2802ccef0f87c1263c6cc2d2fdf6
Summary: Integrate aot flow with model exporter.
Test Plan:
buck test dper3/dper3_backend/delivery/tests:dper3_model_export_test
replayer test see D23407733
Reviewed By: ipiszy
Differential Revision: D23313689
fbshipit-source-id: 39ae8d578ed28ddd6510db959b65974a5ff62888
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/43938
resubmit
Test Plan: unit test included
Reviewed By: mruberry
Differential Revision: D23443493
fbshipit-source-id: 7b68f8f7d1be58bee2154e9a498b5b6a09d11670
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/43591
100 randomized inputs vs 50 doesn't change the balance that much but speed up test runtime
Test Plan: CI
Reviewed By: orionr, seemethere
Differential Revision: D23332393
fbshipit-source-id: 7a8ff9127ee3e045a83658a7a670a844f3862987
Summary:
Separate user embeddings and ad embeddings in blobsOrder. New order:
1. meta_net_def
2. preload_blobs
3. user_embeddings (embeddings in remote request only net)
4. ad_embeddings (embeddings in remote other net)
Add a field requestOnlyEmbeddings in meta_net_def to record user_embeddings.
This is for flash verification.
Test Plan:
buck test dper3/dper3_backend/delivery/tests:blob_reorder_test
Run a flow with canary package f211282476
Check the net: n326826, request_only_embeddings are recorded as expected
Reviewed By: ipiszy
Differential Revision: D23008305
fbshipit-source-id: 9360ba3d078f205832821005e8f151b8314f0cf2
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/43205
A number of tests that forward to `TestLoadSaveBase.load_save` are all marked as flaky due to them regularly taking much longer to start up than hypothesis' default timeout of 200ms. This diff fixes the problem by removing the timeout for `load_save`. This is alright as these tests aren't meant to be testing the performance of these operators.
I would set the deadline to 60s if I could however it appears the that caffe2 github CI uses a different version of hypothesis that doesn't allow using `dateutil.timedelta` so instead of trying to figure out an approach that works on both I've just removed the deadline time.
I've also tagged all existing tasks WRT these failures.
Differential Revision: D23175752
fbshipit-source-id: 324f9ff034df1ac4874797f04f50067149a6ba48
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/42927
added fp16 fusion to net transforms
refactored the transforms as well as glow_transform to get out of opt/custom so that the OSS builds passed
Test Plan: added net runner tests for this
Reviewed By: yinghai
Differential Revision: D23080881
fbshipit-source-id: ee6451811fedfd07c6560c178229854bca29301f
Summary:
1. Fix illegal memory access issue for SplitByLengths operator in the CUDA context.
2. Add support to scaling lengths vector for SplitByLengths operator.
3. Add support to test SplitByLengths operator in the CUDA context.
Example for SplitByLengths operator processing scaling lengths vector:
value vector A = [1, 2, 3, 4, 5, 6]
length vector B = [1, 2]
after execution of SplitByLengths operator,
the output should be [1,2] and [3,4,5,6]
Test Plan: buck test mode/dev-nosan caffe2/caffe2/python/operator_test:concat_split_op_test
Reviewed By: kennyhorror
Differential Revision: D23079841
fbshipit-source-id: 3700e7f2ee0a5a2791850071fdc16e5b054f8400
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/42763
add the fp16 fusions as net transforms:
-layernorm fused with mul+add
-swish int8
Test Plan: added unit test, ran flows
Reviewed By: yinghai
Differential Revision: D23002043
fbshipit-source-id: f0b13d51d68c240b05d2a237a7fb8273e996328b
Summary:
Enforce counter value to double type in rowwise_counter.
**Context:**
The existing implementation is using float type for counter value. But due to the precision limit of a floating number [1], we observed that the counter value can't increment beyond 16777216.0 (i.e., the max value is 16777216.0) in our earlier experiments. We decide to enforce double type to avoid this issue.
[1] https://stackoverflow.com/questions/12596695/why-does-a-float-variable-stop-incrementing-at-16777216-in-c
Test Plan:
op test
```
ruixliu@devvm1997:~/fbsource/fbcode/caffe2/caffe2/python/operator_test(f0b0b48c)$ buck test :rowwise_counter_test
Trace available for this run at /tmp/testpilot.20200728-083200.729292.log
TestPilot test runner for Facebook. See https://fburl.com/testpilot for details.
Testpilot build revision cd2638f1f47250eac058b8c36561760027d16add fbpkg f88726c8ebde4ba288e1172a348c7f46 at Mon Jul 27 18:11:43 2020 by twsvcscm from /usr/local/fbprojects/packages/testinfra.testpilot/887/t.par
Discovering tests
Running 1 test
Started new test run: https://our.intern.facebook.com/intern/testinfra/testrun/7881299364977047
✓ caffe2/caffe2/python/operator_test:rowwise_counter_test - test_rowwise_counter (caffe2.caffe2.python.operator_test.rowwise_counter_test.TestRowWiseCounter) 0.265 1/1 (passed)
✓ caffe2/caffe2/python/operator_test:rowwise_counter_test - main 14.414 (passed)
Finished test run: https://our.intern.facebook.com/intern/testinfra/testrun/7881299364977047
Summary (total time 18.51s):
PASS: 2
FAIL: 0
SKIP: 0
FATAL: 0
TIMEOUT: 0
OMIT: 0
```
optimizer test
```
ruixliu@devvm1997:~/fbsource/fbcode/caffe2/caffe2/python(7d66fbb9)$ buck test :optimizer_test
Finished test run: https://our.intern.facebook.com/intern/testinfra/testrun/7036874434841896
Summary (total time 64.87s):
PASS: 48
FAIL: 0
SKIP: 24
caffe2/caffe2/python:optimizer_test - testGPUDense (caffe2.caffe2.python.optimizer_test.TestMomentumSgd)
caffe2/caffe2/python:optimizer_test - testGPUDense (caffe2.caffe2.python.optimizer_test.TestGFtrl)
caffe2/caffe2/python:optimizer_test - test_caffe2_cpu_vs_numpy (caffe2.caffe2.python.optimizer_test.TestYellowFin)
caffe2/caffe2/python:optimizer_test - testGPUDense (caffe2.caffe2.python.optimizer_test.TestSparseRAdam)
caffe2/caffe2/python:optimizer_test - testGPUDense (caffe2.caffe2.python.optimizer_test.TestRowWiseAdagradWithCounter)
caffe2/caffe2/python:optimizer_test - testGPUDense (caffe2.caffe2.python.optimizer_test.TestAdagrad)
caffe2/caffe2/python:optimizer_test - test_caffe2_gpu_vs_numpy (caffe2.caffe2.python.optimizer_test.TestYellowFin)
caffe2/caffe2/python:optimizer_test - testDense (caffe2.caffe2.python.optimizer_test.TestRowWiseAdagrad)
caffe2/caffe2/python:optimizer_test - testGPUDense (caffe2.caffe2.python.optimizer_test.TestFtrl)
caffe2/caffe2/python:optimizer_test - testSparse (caffe2.caffe2.python.optimizer_test.TestRmsProp)
...and 14 more not shown...
FATAL: 0
TIMEOUT: 0
OMIT: 0
```
param download test
```
ruixliu@devvm1997:~/fbsource/fbcode/caffe2/caffe2/fb/net_transforms/tests(7ef20a38)$ sudo buck test :param_download_test
Finished test run: Finished test run: https://our.intern.facebook.com/intern/testinfra/testrun/6473924481526935
```
e2e flow:
f208394929
f207991149
f207967273
ANP notebook to check the counter value loaded from the flows
https://fburl.com/anp/5fdcbnoi
screenshot of the loaded counter (note that counter max is larger than 16777216.0)
{F250926501}
Reviewed By: ellie-wen
Differential Revision: D22711514
fbshipit-source-id: 426fed7415270aa3f276dda8141907534734337f
Summary:
1. Fix illegal memory access issue for SplitByLengths operator in the CUDA context.
2. Add support to scaling lengths vector for SplitByLengths operator.
3. Add support to test SplitByLengths operator in the CUDA context.
Example for SplitByLengths operator processing scaling lengths vector:
value vector A = [1, 2, 3, 4, 5, 6]
length vector B = [1, 2]
after execution of SplitByLengths operator,
the output should be [1,2] and [3,4,5,6]
Test Plan: buck test mode/dev-nosan caffe2/caffe2/python/operator_test:concat_split_op_test
Reviewed By: kennyhorror
Differential Revision: D22780307
fbshipit-source-id: c5ca60ae16b24032cedfa045a421503b713daa6c
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/42249
Main change is to bring Caffe2's superior error messages for cuda initialization into c10 and use them in all code paths.
Basic logic:
| Case | Call to device_count() | init_cuda, e.g. allocating tensor |
| -- | -- | -- |
| all good | non-zero | just works |
| no gpus | 0, no warning | throw exception with good message |
| driver issues | 0, produce warning | throw exception with good message |
| out of memory with ASAN | 0, produce warning| throw exception with ASAN message |
Previously, the error thrown from init_cuda was very generic and the ASAN warning (if any) was buried in the logs.
Other clean up changes:
* cache device_count() always in a static variable
* move all asan macros in c10
Test Plan:
Hard to unittest because of build modes. Verified manually that the behavior from the table above holds by running the following script in different modes (ASAN/no-ASAN, CUDA_VISIBLE_DEVICES=):
```
print('before import')
import torch
print('after import')
print('devices: ', torch.cuda.device_count())
x = torch.tensor([1,2,3])
print('tensor creation')
x = x.cuda()
print('moved to cuda')
```
Reviewed By: ngimel
Differential Revision: D22824329
fbshipit-source-id: 5314007313a3897fc955b02f8b21b661ae35fdf5
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/42516
att. We need it for some scripts.
Reviewed By: houseroad
Differential Revision: D22918112
fbshipit-source-id: 8a1696ceeeda67a34114bc57cb52c925711cfb4c
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/42421
Previously, we can only feed shape info from Python with float dtype, and batch based dim type when we do onnxifi from Python. This diff removes this limitation and uses TensorBoundShapes protobuf as a generic shape info struct. This will make the onnxifi interface in Python more flexible.
Reviewed By: ChunliF
Differential Revision: D22889781
fbshipit-source-id: 1a89f3a68c215a0409738c425b4e0d0617d58245
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/42381
Introduce new tag to support distributed hogwild.
Reviewed By: boryiingsu
Differential Revision: D20484099
fbshipit-source-id: 5973495589e0a7ab185d3867b37437aa747f408a
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/42380
[Caffe2] Remove explicitly divide by zero in SpatialBN training mode
Test Plan: buck test mode/dev-nosan //caffe2/caffe2/python/operator_test:spatial_bn_op_test
Reviewed By: houseroad
Differential Revision: D22873214
fbshipit-source-id: 70b505391b5db02b45fc46ecd7feb303e50c6280
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/42219
Introduce a new extra info that is tagged on the forward net for the operators sharing the same input. The effect is that the auto gen sum of gradient for the input will not follow the tag of the operator tags in the forward net. This allow more flexible device allocation.
Test Plan:
# unit test
`./buck-out/gen/caffe2/caffe2/python/core_gradients_test#binary.par -r testMultiUseInputAutoGenSumDevice`
Reviewed By: xianjiec, boryiingsu
Differential Revision: D22609080
fbshipit-source-id: d558145e5eb36295580a70e1ee3a822504dd439a
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/42151
Previously our Caffe2 SpatialBN op impl was incorrect for computing running_var without unbias coefficent. Actually it should fail the test because the output will be different with CuDNN's output. However, our tests are too weak to find this bug. This diff fix all of them.
Test Plan: buck test mode/dev-nosan //caffe2/caffe2/python/operator_test:spatial_bn_op_test
Reviewed By: houseroad
Differential Revision: D22786127
fbshipit-source-id: db80becb67d60c44faae180c7e4257cb136a266d
Summary:
Found while trying to get RocM Caffe2 CI green
Pull Request resolved: https://github.com/pytorch/pytorch/pull/42168
Reviewed By: seemethere
Differential Revision: D22791879
Pulled By: malfet
fbshipit-source-id: 8f7ef9711bdc5941b2836e4c8943bb95c72ef8af
Summary:
Found while trying to get RocM Caffe2 job green
Pull Request resolved: https://github.com/pytorch/pytorch/pull/42169
Reviewed By: seemethere
Differential Revision: D22791896
Pulled By: malfet
fbshipit-source-id: 9df6233876aec5ead056365499bab970aa7e8bdc
Summary: we need this op to avoid the splicing of a dense tensor and then use the Mergesinglescaler op
Test Plan: integrated test with dper2
Differential Revision: D22677523
fbshipit-source-id: f4f9a1f06841b0906ec8cbb435482ae0a89e1721
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/41482
This adds a new tag for use with pipeline parallelism.
Test Plan: CI
Reviewed By: heslami
Differential Revision: D22551487
fbshipit-source-id: 90910f458a9bce68f7ef684773322a49aa24494a
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/41687
Specifically, this makes a new library (lazy), which can be used from both core
and workspace.
This allows workspace.Createnet to trigger lazy loading of dyndep dependencies.
Test Plan: Added a unit test specifically for workspace.CreateNet
Reviewed By: dzhulgakov
Differential Revision: D22441877
fbshipit-source-id: 3a9d1af9962585d08ea2566c9c85bec7377d39f2
Summary:
## TLDR
Support using NaN default value for missing dense features in RawInputProcessor for DPER2. In preparation for subsequent support for null flag features in compute meta. For train_eval this is already supported in DPER3 and we do not plan to support this in DPER2 train eval.
Differential Revision: D22439142
fbshipit-source-id: 99ae9755bd41a5d5f43bf5a9a2819d64f3883005
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/41343
Currently caffe2.InitOpLibrary does the dll import uniliaterally. Instead if we make a lazy version and use it, then many pieces of code which do not need the caffe2urrenoperators get a lot faster.
One a real test, the import time went from 140s to 68s. 8s.
This also cleans up the algorithm slightly (although it makes a very minimal
difference), by parsing the list of operators once, rather than every time a
new operator is added, since we defer the RefreshCall until after we've
imported all the operators.
The key way we maintain safety, is that as soon as someone does an operation
which requires a operator (or could), we force importing of all available
operators.
Future work could include trying to identify which code is needed for which
operator and only import the needed ones. There may also be wins available by
playing with dlmopen (which opens within a namespace), or seeing if the dl
flags have an impact (I tried this and didn't see an impact, but dlmopen may
make it better).
Note that this was previously landed and reverted. The issue was that if a import failed and raised an exception, the specific library would not be removed from the lazy imports. This caused our tests which had libraries that failed to poison all other tests that ran after it. This has been fixed and a unit test has been added for this case (to help make it obvious what failed).
Test Plan:
I added a new test a lazy_dyndep_test.py (copied from all_compare_test.py).
I'm a little concerned that I don't see any explicit tests for dyndep, but this
should provide decent coverage.
I've added a specific test to handle the poisoning issues mentioned above, which caused the previous version to get reverted.
Differential Revision: D22506369
fbshipit-source-id: 7395df4778e8eb0220630c570360b99a7d60eb83
Summary: Adding shape inference for SpraseToDense. Proposal impl of shape inference only works when data_to_infer_dim is given, otherwise SpraseToDense output dimension depends on max value of input tensor
Test Plan:
buck test //caffe2/caffe2/python:sparse_to_dense_test
buck test //caffe2/caffe2/python:hypothesis_test -- test_sparse_to_dense
Dper3 Changes:
f204594813
buck test dper3/dper3_models/ads_ranking/model_impl/sparse_nn/tests:sparse_nn_lib_test
Reviewed By: zhongyx12, ChunliF
Differential Revision: D22479511
fbshipit-source-id: 8983a9baea8853deec53ad6f795c874c3fb93de0
Summary:
nccl tests and parallelize_bmuf_distributed test are failing on rocm3.5.1. Skipping these tests to upgrade the CI to rocm3.5.1
jeffdaily sunway513
Pull Request resolved: https://github.com/pytorch/pytorch/pull/41409
Reviewed By: orionr
Differential Revision: D22528928
Pulled By: seemethere
fbshipit-source-id: 928196b7a62a441d391e69f54b278313ecc75d77
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/41313
This diff backs out the backout diff. The failure was due to C++ `or`
not being supported in MSVC. This is now replaced with ||
Original commit changeset: fc7f3f8c968d
Test Plan: Existing unit tests, check github CI.
Reviewed By: malfet
Differential Revision: D22494777
fbshipit-source-id: 3271288919dc3a6bfb82508ab9d021edc910ae45
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/40875
This op uses the given num_bins and a spacing strategy to automatically bin and compute the histogram of given matrices.
Test Plan: Unit tests.
Reviewed By: neha26shah
Differential Revision: D22329069
fbshipit-source-id: 28406b94e284d52d875f73662fc82f93dbc00064
Summary:
unique op test failure in caffe2 blocks upgrading CI to rocm3.5.1. Skipping the test to unblock will re-enable after root causing and fixing the issue.
jeffdaily sunway513
Pull Request resolved: https://github.com/pytorch/pytorch/pull/41219
Differential Revision: D22471452
Pulled By: xw285cornell
fbshipit-source-id: 9e503c8b37c0a4b92632f77b2f8a90281a9889c3
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/39488
Currently caffe2.InitOpLibrary does the dll import uniliaterally. Instead if we make a lazy version and use it, then many pieces of code which do not need the caffe2urrenoperators get a lot faster.
One a real test, the import time went from 140s to 68s. 8s.
This also cleans up the algorithm slightly (although it makes a very minimal
difference), by parsing the list of operators once, rather than every time a
new operator is added, since we defer the RefreshCall until after we've
imported all the operators.
The key way we maintain safety, is that as soon as someone does an operation
which requires a operator (or could), we force importing of all available
operators.
Future work could include trying to identify which code is needed for which
operator and only import the needed ones. There may also be wins available by
playing with dlmopen (which opens within a namespace), or seeing if the dl
flags have an impact (I tried this and didn't see an impact, but dlmopen may
make it better).
Test Plan:
I added a new test a lazy_dyndep_test.py (copied from all_compare_test.py).
I'm a little concerned that I don't see any explicit tests for dyndep, but this
should provide decent coverage.
Differential Revision: D21870844
fbshipit-source-id: 3f65fedb65bb48663670349cee5e1d3e22d560ed
Summary:
This PR contains the following updates:
1. MIOpen 3D pooling enabled in Caffe2.
2. Refactored the MIOpen pooling code in caffe2.
3. Enabled unit test cases for 3D pooling.
CC: ezyang jeffdaily ashishfarmer
Pull Request resolved: https://github.com/pytorch/pytorch/pull/38260
Differential Revision: D21524754
Pulled By: xw285cornell
fbshipit-source-id: ddfe09dc585cd61e42eee22eff8348d326fd0c3b
Summary: Export logit op to pt for better preproc perf
Test Plan:
unit test
Also tested with model re-generation
Reviewed By: houseroad
Differential Revision: D22324611
fbshipit-source-id: 86accb6b4528e5c818d2c3f8c67926f279d158d6
Summary:
Original commit changeset: 46c59d849fa8
The original commit is breaking DPER3 release pipeline with the following failures:
https://www.internalfb.com/intern/chronos/jobinstance?jobinstanceid=9007207344413239&smc=chronos_gp_admin_client&offset=0
```
Child workflow f 202599639 failed with error: c10::Error: [enforce fail at operator.cc:76] blob != nullptr. op Save: Encountered a non-existing input blob: feature_preproc/feature_sparse_to_dense/default_float_value
```
https://www.internalfb.com/intern/chronos/jobinstance?jobinstanceid=9007207344855973&smc=chronos_gp_admin_client&offset=0
```
Child workflow f 202629391 failed with error: c10::Error: [enforce fail at operator.cc:76] blob != nullptr. op Save: Encountered a non-existing input blob: tum_preproc/inductive/feature_sparse_to_dense/default_float_value
```
Related UBN tasks: T69529846, T68986110
Test Plan: Build a DPER3 package on top of this commit, and check that DPER3 release test `model_deliverability_test` is passing.
Differential Revision: D22396317
fbshipit-source-id: 92d5b30cc146c005d6159a8d5bfe8973e2c546dd
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/40856
Add a new activation function - Mish: A Self Regularized Non-Monotonic Neural Activation Function https://arxiv.org/abs/1908.08681
Test Plan:
buck test //caffe2/caffe2/python/operator_test:elementwise_ops_test -- 'test_mish'
{F242275183}
Differential Revision: D22158035
fbshipit-source-id: 459c1dd0ac5b515913fc09b5f4cd13dcf095af31
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/40925
normalization operator does not handle empty tensors correctly. This is a fix.
Test Plan: unit tests
Differential Revision: D22330340
fbshipit-source-id: 0bccf925bb768ebb997ed0c88130c5556308087f
Summary:
## TLDR
Support using NaN default value for missing dense features in RawInputProcessor for DPER2. In preparation for subsequent support for null flag features in compute meta. For train_eval this is already supported in DPER3 and we do not plan to support this in DPER2 train eval.
## Overview
Intern project plan to support adding dense flags for missing feature values instead of replacing with zero.
## Project plan :
https://docs.google.com/document/d/1OsPUTjpJycwxWLCue3Tnb1mx0uDC_2KKWvC1Rwpo2NI/edit?usp=sharing
## Code paths:
See https://fb.quip.com/eFXUA0tbDmNw for the call stack for all affected code paths.
Test Plan:
## fblearner flow test
1. `flow-cli clone f197867430 --run-as-secure-group ads_personalization_systems --force-build` to build a ephemeral package and start a fblearner flow run (may fail)
2. Clone the new run and change the secure_group to `XXXX` and entitlement to `default` in the UI
3. Adds explicit_null_min_coverage flag
4. Optionally reduce `max_examples` since we only test pass/fail instead of quality.
5. Submit the run to test the change
Example:
f198538878
## compare output coverages to daiquery runs
1. Randomly select null flag features from compute meta workflow output
2. Look up the feature id in feature metadata using feature name
3. Check against a daiquery sample of coverage to see if the coverage falls within guidelines.
https://www.internalfb.com/intern/daiquery/workspace/275342740223489/192619942076136/
## Sampled features:
GFF_C66_ADS_USER_SUM_84_PAGE_TYPE_RATIO_EVENT_LIKE_IMPRESSION: 15694257
- original feature compute meta coverage: 0.999992
- daiquery feature coverage (10k rows): 0.69588
- null flag compute meta coverage: 0.293409
GFF_R1303_ADS_USER_SUM_7_PAGE_TYPE_COUNTER_CONVERSION: 16051183
- original feature compute meta coverage: 0.949868
- daiquery feature coverage: 0.82241
- null flag compute meta coverage: 0.151687
## Unit tests:
`buck test fblearner/flow/projects/dper/tests/workflows:ads_test`
https://www.internalfb.com/intern/testinfra/testconsole/testrun/6192449504303863/
Differential Revision: D22026450
fbshipit-source-id: 46c59d849fa89253f14dc2b035c4c677cd6e3a4c
Summary: Use the newly added counter op in sparse adagrad
Reviewed By: chocjy, ellie-wen
Differential Revision: D19221100
fbshipit-source-id: d939d83e3b5b3179f57194be2e8864d0fbbee2c1
Summary:
## TLDR
Support using NaN default value for missing dense features in RawInputProcessor for *DPER2*. In preparation for subsequent support for null flag features in *compute meta*. For train_eval this is already supported in DPER3 and we do not plan to support this in DPER2 train eval.
## Overview
Intern project plan to support adding dense flags for missing feature values instead of replacing with zero.
Project plan :
https://docs.google.com/document/d/1OsPUTjpJycwxWLCue3Tnb1mx0uDC_2KKWvC1Rwpo2NI/edit?usp=sharing
## Code paths:
See https://fb.quip.com/eFXUA0tbDmNw for the call stack for all affected code paths.
Test Plan:
# A. DPER3 blob value inspection
## 1. Build local bento kernel in fbcode folder
`buck build mode/dev-nosan //bento/kernels:bento_kernel_ads_ranking`
## 2. Use kernel `ads_ranking (local)` to print dense feature blob values
n280239
## 2.1 Try `default_dense_value = "0.0"` (default)
```
preproc_6/feature_preproc_6/dper_feature_processor_7/raw_input_proc_7/float_feature_sparse_to_dense_7/float_features [[0. ]
[0. ]
[0. ]
[0. ]
[0. ]
[0. ]
[0. ]
[1. ]
[1.7857143]
[1.7777778]
[1. ]
[0. ]
[0.5625 ]
[0. ]
[0. ]
[0.8 ]
[0. ]
[1. ]
[0.56 ]
[0. ]]
```
## 2.2 Try `default_dense_value = "123"`
```
preproc_2/feature_preproc_2/dper_feature_processor_3/raw_input_proc_3/float_feature_sparse_to_dense_3/float_features [[123. ]
[123. ]
[123. ]
[123. ]
[123. ]
[123. ]
[123. ]
[ 1. ]
[ 1.7857143]
[ 1.7777778]
[ 1. ]
[123. ]
[ 0.5625 ]
[123. ]
[123. ]
[ 0.8 ]
[123. ]
[ 1. ]
[ 0.56 ]
[123. ]]
```
## 2.3 Try `default_dense_value = float("nan")`
```
RuntimeError: [enforce fail at enforce_finite_op.h:40] std::isfinite(input_data[i]). Index 0 is not finite (e.g., NaN, Inf): -nan (Error from operator:
input: "unary_4/logistic_regression_loss_4/average_loss_4/average_loss" name: "" type: "EnforceFinite" device_option { random_seed: 54 })
```
which is expected due to nan input.
# B. Unit test
`buck test fblearner/flow/projects/dper/tests/preprocs:raw_feature_extractor_test`
https://www.internalfb.com/intern/testinfra/testconsole/testrun/5348024586274923/
{F241336814}
Differential Revision: D21961595
fbshipit-source-id: 3dcb153b3c7f42f391584f5e7f52f3d9c76de31f
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/40379
The current sum operator doesn't support Long .. hence modify the code
Test Plan: Write a test case
Reviewed By: jspark1105, yinghai
Differential Revision: D21917365
fbshipit-source-id: b37d2c100c70d17d2f89c309e40360ddfab584ee
Summary:
Pull Request resolved: https://github.com/pytorch/FBGEMM/pull/387
Pull Request resolved: https://github.com/pytorch/pytorch/pull/39985
avx2 optimized 2/4-bit row-wise quantization/dequantization in perfkernels.
This diff slightly change the numerics of quantization by multiplying with the inverse of scale instead of dividing with scale.
Test Plan:
In my devserver
for i in 2 4 8; do echo $i; buck run mode/opt :fused_rowwise_nbit_conversion_bench -- --bit-rate=$i; done
Before this diff
2-bit
3.35394 ms. 100%. FloatToFused2BitRowwiseQuantized
4-bit
3.60351 ms. 100%. FloatToFused4BitRowwiseQuantized
8-bit
0.434467 ms. 100%. FloatToFused8BitRowwiseQuantized
After this diff
2-bit
0.606386 ms. 100%. FloatToFused2BitRowwiseQuantized
4-bit
0.446683 ms. 100%. FloatToFused4BitRowwiseQuantized
8-bit
0.4349 ms. 100%. FloatToFused8BitRowwiseQuantized
Reviewed By: choudharydhruv, jianyuh
Differential Revision: D22033195
fbshipit-source-id: d3a219e47b8345268d90a160c9314ed0d5b71467
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/40081
Adding the functionality to enable timeout of OnnxifiOp run. In the case of backend hanging, it can error out quickly.
Test Plan:
```
buck test glow/fb/test:test_onnxifinnpi -- test_timeout
```
Reviewed By: jackm321
Differential Revision: D22064533
fbshipit-source-id: 25487287c10ab217eb95692f09d48e13e19436ab
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/39407
- support passing a single element tensor as k for topk module
- support passing a single element tensor to constant fill output
Test Plan:
buck test dper3/dper3/modules/tests:core_modules_test -- test_topk_gating_without_split_examples_tensor_k
buck test caffe2/caffe2/python:hypothesis_test -- test_constant_fill_from_tensor
Reviewed By: huayuli00
Differential Revision: D21843739
fbshipit-source-id: 0c5f5c03e9f57eeba40c0068784625164c2527ec
Summary:
Changes in PR https://github.com/pytorch/pytorch/issues/39759 broke HIP caffe2.
hipify for caffe2 renames CUDA to HIP; torch does not.
If caffe2 calls into torch, it needs to use CUDA-named functions.
CC ezyang xw285cornell sunway513 houseroad dzhulgakov
Pull Request resolved: https://github.com/pytorch/pytorch/pull/39801
Differential Revision: D21982493
Pulled By: xw285cornell
fbshipit-source-id: 8e88e0fb80c71f0342e23ef0214a42d5542bdc70
Summary:
THCAllocator functionality is pretty obscure and it's hard to get it working with HIP because of how Caffe2/PyTorch rules are set up (see https://github.com/pytorch/pytorch/issues/39801). Let's just disable the test.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/39843
Reviewed By: zou3519
Differential Revision: D21998687
Pulled By: dzhulgakov
fbshipit-source-id: cd12ba30cdfee658b98393ed3a72e83f4ecf1c9c
Summary:
# Motivations
As explained in the [link](https://stats.stackexchange.com/questions/86991/reason-for-not-shrinking-the-bias-intercept-term-in-regression/161689#161689), regularizing biases will cause mis-calibration of predicted probabilities.
In SparseNN, the unary processor may use 1d embedding tables for the sparse features to serve as biases.
In this diff, the regularization term is automatically skipped for the 1d sparse parameters to avoid regularizing biases.
# Experiments
Experiments were conducted to verify that it has no significant impact on the NE to skip the regularization on 1d sparse parameters.
Baseline.1 (no L2 regularization): f193105372
Baseline.2 (L2 regularization in prod): f193105522
Treatment (skipping L2 regularization on 1d sparse params): f193105708
{F239859690}
Test Plan:
Experiments were conducted to verify that it has no significant impact on the NE to skip the regularization on 1d sparse parameters using a canary package: `aml.dper2.canary:9efc576b35b24361bb600dcbf94d31ea`.
Baseline.1 (no L2 regularization): f193105372
Baseline.2 (L2 regularization in prod): f193105522
Treatment (skipping L2 regularization on 1d sparse params): f193105708
Reviewed By: zhongyx12
Differential Revision: D21757902
fbshipit-source-id: ced126e1eab270669b9981c9ecc287dfc9dee995
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/39373
Line 114 is the only actual change. Other changes are just formatting.
Test Plan: CI
Reviewed By: zrphercule
Differential Revision: D21830893
fbshipit-source-id: 83e49b1b3c48f6bc6de3c48ccce60c84aa49339b
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/39759
Caffe2 has a mode where it uses PT's caching allocator. Somehow we were not calling the initialization explicitly.
Now, I have no idea why it worked before. Probably worth to run a bisect separately.
Reviewed By: houseroad
Differential Revision: D21962331
fbshipit-source-id: f16ad6b27a67dbe0bda93939cca8c94620d22a09
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/38582
Adding LpNorm regularization for sparse features in DPER3. This is done using a sparse regularization op with run_after_optimizer (see D21003029).
* Added code calling new caffe2 operator from D21003029 to caffe2/python/regularizer.py
* Added l1norm and l2norm to sparse regularizer thrift definition.
* Added the new regularization references to test utils.
* Added a new file for unit tests "sparse_nn_sparse_reg_test.py"
Test Plan:
buck test mode/dev //caffe2/caffe2/fb/dper/layer_models/tests:sparse_nn_sparse_reg_test
buck test mode/dev //caffe2/caffe2/fb/dper/layer_models/tests:sparse_nn_reg_test
DPER canary: https://fburl.com/fblearner/rcp5yzeh
New DPER canary: https://fburl.com/fblearner/0krgd74x
Differential Revision: D20704248
fbshipit-source-id: 7e3d5013b3ff3da95ea027f0f2dd855f3ae8e41d
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/39372
we only bump the submodule in oss to unblock some works
Test Plan: ci
Reviewed By: hl475
Differential Revision: D21830800
fbshipit-source-id: fb4a716992efcd71926f7bba24a7c24422c17e38
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/38574
Adding sparse L1 and L2 regularization operator to Caffe2. This doesn't work using run_on_loss, only run_after_optimize. Applying it to run_after_optimize rather than run_on_loss was easier to implement, particularly for the L1 norm which is preferable in some cases and is non-differentiable at zero.
Test Plan: Wrote and ran unit tests in operator_test:sparse_lp_regularizer_test.
Differential Revision: D21003029
fbshipit-source-id: 81070a621752560ce03e320d065ce27807a5d278
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/39297
histogram op doesn't have GPU implementation. It's breaking the CI GPU test. Make the test run cpu only.
Test Plan: CI
Reviewed By: hwangjeff
Differential Revision: D21800824
fbshipit-source-id: 9c835786f22bac7d420ce610397a6ee69084c19a
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/38514
this diff introduces the `Histogram` caffe2 op, which computes a histogram tensor for a list of input tensors. the bin edges of the histogram are defined by arg `bin_edges`.
Test Plan: tests
Reviewed By: chocjy
Differential Revision: D21553956
fbshipit-source-id: fc98c8db691d66d2dad57b6ad14867109913cb6f
Summary:
Previously we got a CI issue in original submission (D21562485), so we backout the original diff (D21588831). Resubmitting here to reprod the CI issue and ask caffe2 dev to take a look.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/38566
Original commit changeset: 6dda4b71904d
Test Plan: buck test
Reviewed By: houseroad
Differential Revision: D21589352
fbshipit-source-id: de40ff2884019e14476e31c4c952f24d6e438f5f
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/37727
Check if the file exists locally only for `log_file_db` db_type. Reader files in other `db_type` like `manifold_log_file_db` are excluded from this check.
Test Plan: Verified that files stored in manifold can be loaded using `DBFileReader`.
Reviewed By: hbjerry
Differential Revision: D21329671
fbshipit-source-id: bbc0e88851783ca3f78f7c61bfe84b480c09b5ac
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/38518
as title
Test Plan: buck test
Reviewed By: olittle
Differential Revision: D21562570
fbshipit-source-id: 3a2e8dea3d821a2bdb9f30db25816a2bfa6c5dcf
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/38517
as title
Test Plan: buck test
Reviewed By: olittle
Differential Revision: D21562485
fbshipit-source-id: 573419e5a8dae4121d99d5b72ed3960a92db7a54
Summary: Issue was introduced in D21258652. We need to make sure it compiles with opt mode. We may still have some left over py2 packages. Let's just use some format work with both.
Test Plan: ci
Reviewed By: xush6528
Differential Revision: D21457394
fbshipit-source-id: cde79a0fc6b4feba307bd9d45e1a1d4a42de9263
Summary:
Skip the tests if network is unaccessible and model can not be downloaded
Pull Request resolved: https://github.com/pytorch/pytorch/pull/37972
Differential Revision: D21441996
Pulled By: malfet
fbshipit-source-id: 5ce59764584974aee9195572338ada1fa0351a75
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/37705
Pull Request resolved: https://github.com/pytorch/pytorch/pull/37372
Posted note: [Regularizing SparseNN Against Over-fitting](https://fb.workplace.com/notes/taiqing-wang/regularizing-sparsenn-against-over-fitting/220306075902708/)
**Problem formulation**
L(w) = J(w) + lambda/2 * ||w||^2
J(w) is the empirical loss, and ||w||^2 is the squared L2 norm of the parameters, a.k.a. L2 regularizer.
dL(w)/ dw_i = dJ(w)/dw_i + lambda w_i
dL(w)/ dw_i is the gradient of L(w) w.r.t. w_i.
To implement the L2 regularizer, the gradient of J(w) w.r.t. w_i is added with w_i. lambda is called as weight decay in this implementation.
**Code changes**
* In the initialization method of AdagradOptimizer, a new input argument, weight_decay, is added.
* In the _run function of AdagradOptimizer, the weight decay will be skipped for 1d bias vectors.
* In the parameter update functions of Adagrad, the gradient is updated by weight_decay * w_i. The default value for weight_decay is zero.
Test Plan:
`
buck build caffe2/caffe2/fb/dper/layer_models/tests/split_1:sparse_nn_test_weight_decay
`
`
./buck-out/gen/caffe2/caffe2/fb/dper/layer_models/tests/split_1/sparse_nn_test_weight_decay#binary.par
`
Reviewed By: jspark1105
Differential Revision: D21258652
fbshipit-source-id: d2366ddcd736a03205a2d16f914703b16d9fce8f
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/37591
skip the tests since gluster is gone.
Test Plan: ci
Reviewed By: ezyang
Differential Revision: D21330359
fbshipit-source-id: a4e158fb72eddb08ba49fcfa9541569a150f8481
Summary:
This PR fixes a couple of syntax errors in `torch/` that prevent MyPy from running, fixes simple type annotation errors (e.g. missing `from typing import List, Tuple, Optional`), and adds granular ignores for errors in particular modules as well as for missing typing in third party packages.
As a result, running `mypy` in the root dir of the repo now runs on:
- `torch/`
- `aten/src/ATen/function_wrapper.py` (the only file already covered in CI)
In CI this runs on GitHub Actions, job Lint, sub-job "quick-checks", task "MyPy typecheck". It should give (right now): `Success: no issues found in 329 source files`.
Here are the details of the original 855 errors when running `mypy torch` on current master (after fixing the couple of syntax errors that prevent `mypy` from running through):
<details>
```
torch/utils/tensorboard/_proto_graph.py:1: error: Cannot find implementation or library stub for module named 'tensorboard.compat.proto.node_def_pb2'
torch/utils/tensorboard/_proto_graph.py:2: error: Cannot find implementation or library stub for module named 'tensorboard.compat.proto.attr_value_pb2'
torch/utils/tensorboard/_proto_graph.py:3: error: Cannot find implementation or library stub for module named 'tensorboard.compat.proto.tensor_shape_pb2'
torch/utils/backcompat/__init__.py:1: error: Cannot find implementation or library stub for module named 'torch._C'
torch/for_onnx/__init__.py:1: error: Cannot find implementation or library stub for module named 'torch.for_onnx.onnx'
torch/cuda/nvtx.py:2: error: Cannot find implementation or library stub for module named 'torch._C'
torch/utils/show_pickle.py:59: error: Name 'pickle._Unpickler' is not defined
torch/utils/show_pickle.py:113: error: "Type[PrettyPrinter]" has no attribute "_dispatch"
torch/utils/tensorboard/_onnx_graph.py:1: error: Cannot find implementation or library stub for module named 'tensorboard.compat.proto.graph_pb2'
torch/utils/tensorboard/_onnx_graph.py:2: error: Cannot find implementation or library stub for module named 'tensorboard.compat.proto.node_def_pb2'
torch/utils/tensorboard/_onnx_graph.py:3: error: Cannot find implementation or library stub for module named 'tensorboard.compat.proto.versions_pb2'
torch/utils/tensorboard/_onnx_graph.py:4: error: Cannot find implementation or library stub for module named 'tensorboard.compat.proto.attr_value_pb2'
torch/utils/tensorboard/_onnx_graph.py:5: error: Cannot find implementation or library stub for module named 'tensorboard.compat.proto.tensor_shape_pb2'
torch/utils/tensorboard/_onnx_graph.py:9: error: Cannot find implementation or library stub for module named 'onnx'
torch/contrib/_tensorboard_vis.py:10: error: Cannot find implementation or library stub for module named 'tensorflow.core.util'
torch/contrib/_tensorboard_vis.py:11: error: Cannot find implementation or library stub for module named 'tensorflow.core.framework'
torch/contrib/_tensorboard_vis.py:12: error: Cannot find implementation or library stub for module named 'tensorflow.python.summary.writer.writer'
torch/utils/hipify/hipify_python.py:43: error: Need type annotation for 'CAFFE2_TEMPLATE_MAP' (hint: "CAFFE2_TEMPLATE_MAP: Dict[<type>, <type>] = ...")
torch/utils/hipify/hipify_python.py:636: error: "object" has no attribute "items"
torch/nn/_reduction.py:27: error: Name 'Optional' is not defined
torch/nn/_reduction.py:27: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Optional")
torch/nn/_reduction.py:47: error: Name 'Optional' is not defined
torch/nn/_reduction.py:47: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Optional")
torch/utils/tensorboard/_utils.py:17: error: Skipping analyzing 'matplotlib.pyplot': found module but no type hints or library stubs
torch/utils/tensorboard/_utils.py:17: error: Skipping analyzing 'matplotlib': found module but no type hints or library stubs
torch/utils/tensorboard/_utils.py:18: error: Skipping analyzing 'matplotlib.backends.backend_agg': found module but no type hints or library stubs
torch/utils/tensorboard/_utils.py:18: error: Skipping analyzing 'matplotlib.backends': found module but no type hints or library stubs
torch/nn/modules/utils.py:27: error: Name 'List' is not defined
torch/nn/modules/utils.py:27: note: Did you forget to import it from "typing"? (Suggestion: "from typing import List")
caffe2/proto/caffe2_pb2.py:17: error: Unexpected keyword argument "serialized_options" for "FileDescriptor"; did you mean "serialized_pb"?
caffe2/proto/caffe2_pb2.py:25: error: Unexpected keyword argument "serialized_options" for "EnumDescriptor"
caffe2/proto/caffe2_pb2.py:31: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:35: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:39: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:43: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:47: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:51: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:55: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:59: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:63: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:67: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:71: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:75: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:102: error: Unexpected keyword argument "serialized_options" for "EnumDescriptor"
caffe2/proto/caffe2_pb2.py:108: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:112: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:124: error: Unexpected keyword argument "serialized_options" for "EnumDescriptor"
caffe2/proto/caffe2_pb2.py:130: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:134: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:138: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:142: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:146: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:150: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:154: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:158: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:162: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:166: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:170: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:174: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:178: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:182: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:194: error: Unexpected keyword argument "serialized_options" for "EnumDescriptor"
caffe2/proto/caffe2_pb2.py:200: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:204: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:208: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:212: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:224: error: Unexpected keyword argument "serialized_options" for "EnumDescriptor"
caffe2/proto/caffe2_pb2.py:230: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:234: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:238: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:242: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:246: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:250: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:254: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/caffe2_pb2.py:267: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/caffe2_pb2.py:274: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:281: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:288: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:295: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:302: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:327: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/caffe2_pb2.py:334: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:341: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:364: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/caffe2_pb2.py:371: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:378: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:385: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:392: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:399: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:406: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:413: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:420: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:427: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:434: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:441: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:448: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:455: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:462: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:488: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/caffe2_pb2.py:495: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:502: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:509: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:516: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:523: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:530: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:537: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:544: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:551: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:558: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:565: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:572: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:596: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/caffe2_pb2.py:603: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:627: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/caffe2_pb2.py:634: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:641: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:648: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:655: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:662: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:686: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/caffe2_pb2.py:693: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:717: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/caffe2_pb2.py:724: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:731: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:738: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:763: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/caffe2_pb2.py:770: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:777: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:784: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:808: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/caffe2_pb2.py:815: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:822: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:829: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:836: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:843: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:850: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:857: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:864: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:871: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:878: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:885: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:892: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:916: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/caffe2_pb2.py:923: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:930: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:937: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:944: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:951: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:958: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:982: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/caffe2_pb2.py:989: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:996: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1003: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1010: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1017: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1024: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1031: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1038: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1045: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1052: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1059: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1066: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1090: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/caffe2_pb2.py:1097: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1104: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1128: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/caffe2_pb2.py:1135: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1142: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1166: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/caffe2_pb2.py:1173: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1180: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1187: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1194: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1218: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/caffe2_pb2.py:1225: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1232: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1239: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1246: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1253: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1260: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1267: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1274: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1281: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1305: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/caffe2_pb2.py:1312: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1319: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1326: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1333: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1340: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1347: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1354: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1361: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1368: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1375: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1382: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1389: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1396: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1420: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/caffe2_pb2.py:1427: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1434: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1441: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1465: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/caffe2_pb2.py:1472: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1479: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1486: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1493: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1500: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1507: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1514: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1538: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/caffe2_pb2.py:1545: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1552: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1559: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1566: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/caffe2_pb2.py:1667: error: "GeneratedProtocolMessageType" has no attribute "Segment"
torch/multiprocessing/queue.py:4: error: No library stub file for standard library module 'multiprocessing.reduction'
caffe2/proto/torch_pb2.py:18: error: Unexpected keyword argument "serialized_options" for "FileDescriptor"; did you mean "serialized_pb"?
caffe2/proto/torch_pb2.py:27: error: Unexpected keyword argument "serialized_options" for "EnumDescriptor"
caffe2/proto/torch_pb2.py:33: error: Unexpected keyword argument "serialized_options" for "EnumValueDescriptor"
caffe2/proto/torch_pb2.py:50: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/torch_pb2.py:57: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:81: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/torch_pb2.py:88: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:95: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:102: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:109: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:116: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:123: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:130: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:137: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:144: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:151: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:175: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/torch_pb2.py:182: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:189: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:196: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:220: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/torch_pb2.py:227: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:234: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:241: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:265: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/torch_pb2.py:272: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:279: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:286: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:293: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:300: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:307: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:314: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:321: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:328: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:335: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:342: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:366: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/torch_pb2.py:373: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:397: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/torch_pb2.py:404: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:411: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:418: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:425: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/torch_pb2.py:432: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/metanet_pb2.py:17: error: Unexpected keyword argument "serialized_options" for "FileDescriptor"; did you mean "serialized_pb"?
caffe2/proto/metanet_pb2.py:29: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/metanet_pb2.py:36: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/metanet_pb2.py:43: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/metanet_pb2.py:50: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/metanet_pb2.py:57: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/metanet_pb2.py:64: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/metanet_pb2.py:88: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/metanet_pb2.py:95: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/metanet_pb2.py:102: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/metanet_pb2.py:126: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/metanet_pb2.py:133: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/metanet_pb2.py:140: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/metanet_pb2.py:164: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/metanet_pb2.py:171: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/metanet_pb2.py:178: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/metanet_pb2.py:202: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/metanet_pb2.py:209: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/metanet_pb2.py:216: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/metanet_pb2.py:240: error: Unexpected keyword argument "serialized_options" for "Descriptor"
caffe2/proto/metanet_pb2.py:247: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/metanet_pb2.py:254: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/metanet_pb2.py:261: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/metanet_pb2.py:268: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/metanet_pb2.py:275: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/metanet_pb2.py:282: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/metanet_pb2.py:289: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/metanet_pb2.py:296: error: Unexpected keyword argument "serialized_options" for "FieldDescriptor"
caffe2/proto/__init__.py:13: error: Skipping analyzing 'caffe2.caffe2.fb.session.proto': found module but no type hints or library stubs
torch/multiprocessing/pool.py:3: error: No library stub file for standard library module 'multiprocessing.util'
torch/multiprocessing/pool.py:3: note: (Stub files are from https://github.com/python/typeshed)
caffe2/python/scope.py:10: error: Skipping analyzing 'past.builtins': found module but no type hints or library stubs
caffe2/python/__init__.py:7: error: Module has no attribute "CPU"
caffe2/python/__init__.py:8: error: Module has no attribute "CUDA"
caffe2/python/__init__.py:9: error: Module has no attribute "MKLDNN"
caffe2/python/__init__.py:10: error: Module has no attribute "OPENGL"
caffe2/python/__init__.py:11: error: Module has no attribute "OPENCL"
caffe2/python/__init__.py:12: error: Module has no attribute "IDEEP"
caffe2/python/__init__.py:13: error: Module has no attribute "HIP"
caffe2/python/__init__.py:14: error: Module has no attribute "COMPILE_TIME_MAX_DEVICE_TYPES"; maybe "PROTO_COMPILE_TIME_MAX_DEVICE_TYPES"?
caffe2/python/__init__.py:15: error: Module has no attribute "ONLY_FOR_TEST"; maybe "PROTO_ONLY_FOR_TEST"?
caffe2/python/__init__.py:34: error: Item "_Loader" of "Optional[_Loader]" has no attribute "exec_module"
caffe2/python/__init__.py:34: error: Item "None" of "Optional[_Loader]" has no attribute "exec_module"
caffe2/python/__init__.py:35: error: Module has no attribute "cuda"
caffe2/python/__init__.py:37: error: Module has no attribute "cuda"
caffe2/python/__init__.py:49: error: Module has no attribute "add_dll_directory"
torch/random.py:4: error: Cannot find implementation or library stub for module named 'torch._C'
torch/_classes.py:2: error: Cannot find implementation or library stub for module named 'torch._C'
torch/onnx/__init__.py:1: error: Cannot find implementation or library stub for module named 'torch._C'
torch/hub.py:21: error: Skipping analyzing 'tqdm.auto': found module but no type hints or library stubs
torch/hub.py:24: error: Skipping analyzing 'tqdm': found module but no type hints or library stubs
torch/hub.py:27: error: Name 'tqdm' already defined (possibly by an import)
torch/_tensor_str.py:164: error: Not all arguments converted during string formatting
torch/_ops.py:1: error: Cannot find implementation or library stub for module named 'torch._C'
torch/_linalg_utils.py:26: error: Name 'Optional' is not defined
torch/_linalg_utils.py:26: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Optional")
torch/_linalg_utils.py:26: error: Name 'Tensor' is not defined
torch/_linalg_utils.py:63: error: Name 'Tensor' is not defined
torch/_linalg_utils.py:63: error: Name 'Optional' is not defined
torch/_linalg_utils.py:63: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Optional")
torch/_linalg_utils.py:70: error: Name 'Optional' is not defined
torch/_linalg_utils.py:70: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Optional")
torch/_linalg_utils.py:70: error: Name 'Tensor' is not defined
torch/_linalg_utils.py:88: error: Name 'Tensor' is not defined
torch/_linalg_utils.py:88: error: Name 'Optional' is not defined
torch/_linalg_utils.py:88: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Optional")
torch/_linalg_utils.py:88: error: Name 'Tuple' is not defined
torch/_linalg_utils.py:88: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Tuple")
torch/_jit_internal.py:17: error: Need type annotation for 'boolean_dispatched'
torch/_jit_internal.py:474: error: Need type annotation for '_overloaded_fns' (hint: "_overloaded_fns: Dict[<type>, <type>] = ...")
torch/_jit_internal.py:512: error: Need type annotation for '_overloaded_methods' (hint: "_overloaded_methods: Dict[<type>, <type>] = ...")
torch/_jit_internal.py:648: error: Incompatible types in assignment (expression has type "FinalCls", variable has type "_SpecialForm")
torch/sparse/__init__.py:11: error: Name 'Tensor' is not defined
torch/sparse/__init__.py:71: error: Name 'Tensor' is not defined
torch/sparse/__init__.py:71: error: Name 'Optional' is not defined
torch/sparse/__init__.py:71: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Optional")
torch/sparse/__init__.py:71: error: Name 'Tuple' is not defined
torch/sparse/__init__.py:71: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Tuple")
torch/nn/init.py:109: error: Name 'Tensor' is not defined
torch/nn/init.py:126: error: Name 'Tensor' is not defined
torch/nn/init.py:142: error: Name 'Tensor' is not defined
torch/nn/init.py:165: error: Name 'Tensor' is not defined
torch/nn/init.py:180: error: Name 'Tensor' is not defined
torch/nn/init.py:194: error: Name 'Tensor' is not defined
torch/nn/init.py:287: error: Name 'Tensor' is not defined
torch/nn/init.py:315: error: Name 'Tensor' is not defined
torch/multiprocessing/reductions.py:8: error: No library stub file for standard library module 'multiprocessing.util'
torch/multiprocessing/reductions.py:9: error: No library stub file for standard library module 'multiprocessing.reduction'
torch/multiprocessing/reductions.py:17: error: No library stub file for standard library module 'multiprocessing.resource_sharer'
torch/jit/_builtins.py:72: error: Module has no attribute "_no_grad_embedding_renorm_"
torch/jit/_builtins.py:80: error: Module has no attribute "stft"
torch/jit/_builtins.py:81: error: Module has no attribute "cdist"
torch/jit/_builtins.py:82: error: Module has no attribute "norm"
torch/jit/_builtins.py:83: error: Module has no attribute "nuclear_norm"
torch/jit/_builtins.py:84: error: Module has no attribute "frobenius_norm"
torch/backends/cudnn/__init__.py:8: error: Cannot find implementation or library stub for module named 'torch._C'
torch/backends/cudnn/__init__.py:86: error: Need type annotation for '_handles' (hint: "_handles: Dict[<type>, <type>] = ...")
torch/autograd/profiler.py:13: error: Name 'ContextDecorator' already defined (possibly by an import)
torch/autograd/function.py:2: error: Cannot find implementation or library stub for module named 'torch._C'
torch/autograd/function.py:2: note: See https://mypy.readthedocs.io/en/latest/running_mypy.html#missing-imports
torch/autograd/function.py:109: error: Unsupported dynamic base class "with_metaclass"
torch/serialization.py:609: error: "Callable[[Any], Any]" has no attribute "cache"
torch/_lowrank.py:11: error: Name 'Tensor' is not defined
torch/_lowrank.py:13: error: Name 'Optional' is not defined
torch/_lowrank.py:13: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Optional")
torch/_lowrank.py:14: error: Name 'Optional' is not defined
torch/_lowrank.py:14: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Optional")
torch/_lowrank.py:14: error: Name 'Tensor' is not defined
torch/_lowrank.py:82: error: Name 'Tensor' is not defined
torch/_lowrank.py:82: error: Name 'Optional' is not defined
torch/_lowrank.py:82: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Optional")
torch/_lowrank.py:82: error: Name 'Tuple' is not defined
torch/_lowrank.py:82: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Tuple")
torch/_lowrank.py:130: error: Name 'Tensor' is not defined
torch/_lowrank.py:130: error: Name 'Optional' is not defined
torch/_lowrank.py:130: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Optional")
torch/_lowrank.py:130: error: Name 'Tuple' is not defined
torch/_lowrank.py:130: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Tuple")
torch/_lowrank.py:167: error: Name 'Tensor' is not defined
torch/_lowrank.py:167: error: Name 'Optional' is not defined
torch/_lowrank.py:167: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Optional")
torch/_lowrank.py:167: error: Name 'Tuple' is not defined
torch/_lowrank.py:167: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Tuple")
torch/quantization/observer.py:45: error: Variable "torch.quantization.observer.ABC" is not valid as a type
torch/quantization/observer.py:45: note: See https://mypy.readthedocs.io/en/latest/common_issues.html#variables-vs-type-aliases
torch/quantization/observer.py:45: error: Invalid base class "ABC"
torch/quantization/observer.py:127: error: Name 'Tensor' is not defined
torch/quantization/observer.py:127: error: Name 'Tuple' is not defined
torch/quantization/observer.py:127: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Tuple")
torch/quantization/observer.py:172: error: Module has no attribute "per_tensor_symmetric"
torch/quantization/observer.py:172: error: Module has no attribute "per_channel_symmetric"
torch/quantization/observer.py:192: error: Name 'Tensor' is not defined
torch/quantization/observer.py:192: error: Name 'Tuple' is not defined
torch/quantization/observer.py:192: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Tuple")
torch/quantization/observer.py:233: error: Module has no attribute "per_tensor_symmetric"
torch/quantization/observer.py:233: error: Module has no attribute "per_channel_symmetric"
torch/quantization/observer.py:534: error: Name 'Tensor' is not defined
torch/quantization/observer.py:885: error: Name 'Tensor' is not defined
torch/quantization/observer.py:885: error: Name 'Tuple' is not defined
torch/quantization/observer.py:885: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Tuple")
torch/quantization/observer.py:894: error: Cannot determine type of 'max_val'
torch/quantization/observer.py:894: error: Cannot determine type of 'min_val'
torch/quantization/observer.py:899: error: Cannot determine type of 'min_val'
torch/quantization/observer.py:902: error: Name 'Tensor' is not defined
torch/quantization/observer.py:925: error: Name 'Tensor' is not defined
torch/quantization/observer.py:928: error: Cannot determine type of 'min_val'
torch/quantization/observer.py:929: error: Cannot determine type of 'max_val'
torch/quantization/observer.py:946: error: Argument "min" to "histc" has incompatible type "Tuple[Tensor, Tensor]"; expected "Union[int, float, bool]"
torch/quantization/observer.py:946: error: Argument "max" to "histc" has incompatible type "Tuple[Tensor, Tensor]"; expected "Union[int, float, bool]"
torch/quantization/observer.py:1056: error: Module has no attribute "per_tensor_symmetric"
torch/quantization/observer.py:1058: error: Module has no attribute "per_channel_symmetric"
torch/nn/quantized/functional.py:76: error: Name 'Tensor' is not defined
torch/nn/quantized/functional.py:76: error: Name 'BroadcastingList2' is not defined
torch/nn/quantized/functional.py:259: error: Name 'Tensor' is not defined
torch/nn/quantized/functional.py:259: error: Name 'Optional' is not defined
torch/nn/quantized/functional.py:259: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Optional")
torch/nn/quantized/functional.py:289: error: Module has no attribute "ops"
torch/nn/quantized/functional.py:290: error: Module has no attribute "ops"
torch/nn/quantized/functional.py:308: error: Name 'Tensor' is not defined
torch/nn/quantized/functional.py:326: error: Name 'Tensor' is not defined
torch/nn/quantized/functional.py:356: error: Name 'Tensor' is not defined
torch/nn/quantized/functional.py:371: error: Name 'Tensor' is not defined
torch/nn/quantized/functional.py:400: error: Name 'Tensor' is not defined
torch/nn/quantized/functional.py:400: error: Name 'Optional' is not defined
torch/nn/quantized/functional.py:400: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Optional")
torch/nn/quantized/functional.py:430: error: Name 'Tensor' is not defined
torch/nn/quantized/functional.py:448: error: Name 'Tensor' is not defined
torch/nn/quantized/modules/linear.py:26: error: Module has no attribute "ops"
torch/nn/quantized/modules/linear.py:28: error: Module has no attribute "ops"
torch/nn/quantized/modules/functional_modules.py:40: error: Name 'Tensor' is not defined
torch/nn/quantized/modules/functional_modules.py:47: error: Name 'Tensor' is not defined
torch/nn/quantized/modules/functional_modules.py:54: error: Name 'Tensor' is not defined
torch/nn/quantized/modules/functional_modules.py:61: error: Name 'Tensor' is not defined
torch/nn/quantized/modules/functional_modules.py:68: error: Name 'List' is not defined
torch/nn/quantized/modules/functional_modules.py:68: note: Did you forget to import it from "typing"? (Suggestion: "from typing import List")
torch/nn/quantized/modules/functional_modules.py:68: error: Name 'Tensor' is not defined
torch/nn/quantized/modules/functional_modules.py:75: error: Name 'Tensor' is not defined
torch/nn/quantized/modules/functional_modules.py:140: error: Name 'Tensor' is not defined
torch/nn/quantized/modules/functional_modules.py:146: error: Name 'Tensor' is not defined
torch/nn/quantized/modules/functional_modules.py:151: error: Name 'Tensor' is not defined
torch/nn/quantized/modules/functional_modules.py:157: error: Name 'Tensor' is not defined
torch/nn/quantized/modules/functional_modules.py:162: error: Name 'List' is not defined
torch/nn/quantized/modules/functional_modules.py:162: note: Did you forget to import it from "typing"? (Suggestion: "from typing import List")
torch/nn/quantized/modules/functional_modules.py:162: error: Name 'Tensor' is not defined
torch/nn/quantized/modules/functional_modules.py:168: error: Name 'Tensor' is not defined
torch/multiprocessing/spawn.py:9: error: Module 'torch.multiprocessing' has no attribute '_prctl_pr_set_pdeathsig'
torch/multiprocessing/__init__.py:28: error: Module has no attribute "__all__"
torch/jit/frontend.py:9: error: Cannot find implementation or library stub for module named 'torch._C._jit_tree_views'
torch/jit/annotations.py:6: error: Module 'torch._jit_internal' has no attribute 'BroadcastingList2'; maybe "BroadcastingList1" or "BroadcastingListCls"?
torch/jit/annotations.py:6: error: Module 'torch._jit_internal' has no attribute 'BroadcastingList3'; maybe "BroadcastingList1" or "BroadcastingListCls"?
torch/jit/annotations.py:9: error: Cannot find implementation or library stub for module named 'torch._C'
torch/distributions/distribution.py:16: error: Need type annotation for 'arg_constraints' (hint: "arg_constraints: Dict[<type>, <type>] = ...")
torch/distributions/distribution.py:74: error: Name 'arg_constraints' already defined on line 16
torch/distributions/distribution.py:84: error: Name 'support' already defined on line 15
torch/functional.py:114: error: Name 'Tuple' is not defined
torch/functional.py:114: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Tuple")
torch/functional.py:114: error: Name 'Optional' is not defined
torch/functional.py:114: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Optional")
torch/functional.py:189: error: Incompatible types in assignment (expression has type "None", variable has type "Tensor")
torch/functional.py:200: error: Argument 1 to "_indices_product" has incompatible type "Tuple[int, ...]"; expected "List[int]"
torch/functional.py:204: error: No overload variant of "__setitem__" of "list" matches argument types "Tensor", "int"
torch/functional.py:204: note: Possible overload variants:
torch/functional.py:204: note: def __setitem__(self, int, int) -> None
torch/functional.py:204: note: def __setitem__(self, slice, Iterable[int]) -> None
torch/functional.py:204: error: No overload variant of "__getitem__" of "list" matches argument type "Tensor"
torch/functional.py:204: note: def __getitem__(self, int) -> int
torch/functional.py:204: note: def __getitem__(self, slice) -> List[int]
torch/functional.py:207: error: "Tensor" has no attribute "copy_"
torch/functional.py:212: error: No overload variant of "__setitem__" of "list" matches argument types "Tensor", "int"
torch/functional.py:212: note: Possible overload variants:
torch/functional.py:212: note: def __setitem__(self, int, int) -> None
torch/functional.py:212: note: def __setitem__(self, slice, Iterable[int]) -> None
torch/functional.py:212: error: No overload variant of "__getitem__" of "list" matches argument type "Tensor"
torch/functional.py:212: note: def __getitem__(self, int) -> int
torch/functional.py:212: note: def __getitem__(self, slice) -> List[int]
torch/functional.py:215: error: Incompatible types in assignment (expression has type "None", variable has type "Tensor")
torch/functional.py:334: error: Name 'Optional' is not defined
torch/functional.py:334: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Optional")
torch/functional.py:429: error: Argument 2 to "pad" has incompatible type "Tuple[int, int]"; expected "List[int]"
torch/functional.py:431: error: Module has no attribute "stft"
torch/functional.py:766: error: Module has no attribute "cdist"
torch/functional.py:768: error: Module has no attribute "cdist"
torch/functional.py:770: error: Module has no attribute "cdist"
torch/functional.py:775: error: Name 'Optional' is not defined
torch/functional.py:775: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Optional")
torch/functional.py:780: error: Name 'Optional' is not defined
torch/functional.py:780: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Optional")
torch/functional.py:780: error: Name 'number' is not defined
torch/functional.py:780: error: Name 'norm' already defined on line 775
torch/functional.py:785: error: Name 'Optional' is not defined
torch/functional.py:785: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Optional")
torch/functional.py:785: error: Name 'number' is not defined
torch/functional.py:785: error: Name 'norm' already defined on line 775
torch/functional.py:790: error: Name 'Optional' is not defined
torch/functional.py:790: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Optional")
torch/functional.py:790: error: Name 'norm' already defined on line 775
torch/functional.py:795: error: Name 'norm' already defined on line 775
torch/functional.py:960: error: Name 'Any' is not defined
torch/functional.py:960: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Any")
torch/functional.py:960: error: Name 'Tuple' is not defined
torch/functional.py:960: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Tuple")
torch/functional.py:1036: error: Argument 1 to "len" has incompatible type "int"; expected "Sized"
torch/functional.py:1041: error: Name 'Optional' is not defined
torch/functional.py:1041: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Optional")
torch/functional.py:1041: error: Name 'Tuple' is not defined
torch/functional.py:1041: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Tuple")
torch/functional.py:1056: error: Name 'Optional' is not defined
torch/functional.py:1056: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Optional")
torch/functional.py:1056: error: Name 'Tuple' is not defined
torch/functional.py:1056: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Tuple")
torch/distributions/von_mises.py:87: error: Incompatible types in assignment (expression has type "_Real", base class "Distribution" defined the type as "None")
torch/distributions/negative_binomial.py:25: error: Incompatible types in assignment (expression has type "_IntegerGreaterThan", base class "Distribution" defined the type as "None")
torch/distributions/multivariate_normal.py:116: error: Incompatible types in assignment (expression has type "_Real", base class "Distribution" defined the type as "None")
torch/distributions/laplace.py:23: error: Incompatible types in assignment (expression has type "_Real", base class "Distribution" defined the type as "None")
torch/distributions/independent.py:34: error: Need type annotation for 'arg_constraints' (hint: "arg_constraints: Dict[<type>, <type>] = ...")
torch/distributions/cauchy.py:28: error: Incompatible types in assignment (expression has type "_Real", base class "Distribution" defined the type as "None")
torch/distributions/poisson.py:28: error: Incompatible types in assignment (expression has type "_IntegerGreaterThan", base class "Distribution" defined the type as "None")
torch/distributions/one_hot_categorical.py:32: error: Incompatible types in assignment (expression has type "_Simplex", base class "Distribution" defined the type as "None")
torch/distributions/normal.py:27: error: Incompatible types in assignment (expression has type "_Real", base class "Distribution" defined the type as "None")
torch/distributions/lowrank_multivariate_normal.py:79: error: Incompatible types in assignment (expression has type "_Real", base class "Distribution" defined the type as "None")
torch/distributions/gamma.py:30: error: Incompatible types in assignment (expression has type "_GreaterThan", base class "Distribution" defined the type as "None")
torch/distributions/exponential.py:23: error: Incompatible types in assignment (expression has type "_GreaterThan", base class "Distribution" defined the type as "None")
torch/distributions/fishersnedecor.py:25: error: Incompatible types in assignment (expression has type "_GreaterThan", base class "Distribution" defined the type as "None")
torch/distributions/dirichlet.py:44: error: Incompatible types in assignment (expression has type "_Simplex", base class "Distribution" defined the type as "None")
torch/nn/quantized/dynamic/modules/rnn.py:230: error: Incompatible types in assignment (expression has type "int", variable has type "Tensor")
torch/nn/quantized/dynamic/modules/rnn.py:232: error: Incompatible types in assignment (expression has type "int", variable has type "Tensor")
torch/nn/quantized/dynamic/modules/rnn.py:236: error: Incompatible return value type (got "Tuple[Any, Tensor, Any]", expected "Tuple[int, int, int]")
torch/nn/quantized/dynamic/modules/rnn.py:351: error: Incompatible types in assignment (expression has type "Type[LSTM]", base class "RNNBase" defined the type as "Type[RNNBase]")
torch/nn/quantized/dynamic/modules/rnn.py:381: error: Module has no attribute "quantized_lstm"
torch/nn/quantized/dynamic/modules/rnn.py:385: error: Module has no attribute "quantized_lstm"
torch/nn/quantized/dynamic/modules/rnn.py:414: error: Argument 1 to "forward_impl" of "LSTM" has incompatible type "PackedSequence"; expected "Tensor"
torch/nn/quantized/dynamic/modules/rnn.py:416: error: Incompatible types in assignment (expression has type "PackedSequence", variable has type "Tensor")
torch/nn/quantized/dynamic/modules/rnn.py:418: error: Incompatible return value type (got "Tuple[Tensor, Tuple[Tensor, Tensor]]", expected "Tuple[PackedSequence, Tuple[Tensor, Tensor]]")
torch/nn/quantized/dynamic/modules/rnn.py:420: error: Argument 1 of "permute_hidden" is incompatible with supertype "RNNBase"; supertype defines the argument type as "Tensor"
torch/nn/quantized/dynamic/modules/rnn.py:420: error: Return type "Tuple[Tensor, Tensor]" of "permute_hidden" incompatible with return type "Tensor" in supertype "RNNBase"
torch/nn/quantized/dynamic/modules/rnn.py:426: error: Argument 2 of "check_forward_args" is incompatible with supertype "RNNBase"; supertype defines the argument type as "Tensor"
torch/nn/intrinsic/qat/modules/conv_fused.py:232: error: Incompatible types in assignment (expression has type "Type[ConvBnReLU2d]", base class "ConvBn2d" defined the type as "Type[ConvBn2d]")
torch/distributions/beta.py:27: error: Incompatible types in assignment (expression has type "_Interval", base class "Distribution" defined the type as "None")
torch/distributions/geometric.py:31: error: Incompatible types in assignment (expression has type "_IntegerGreaterThan", base class "Distribution" defined the type as "None")
torch/distributions/continuous_bernoulli.py:38: error: Incompatible types in assignment (expression has type "_Interval", base class "Distribution" defined the type as "None")
torch/distributions/bernoulli.py:30: error: Incompatible types in assignment (expression has type "_Boolean", base class "Distribution" defined the type as "None")
torch/quantization/fake_quantize.py:126: error: Module has no attribute "per_tensor_symmetric"
torch/quantization/fake_quantize.py:132: error: Module has no attribute "per_channel_symmetric"
torch/distributions/transformed_distribution.py:41: error: Need type annotation for 'arg_constraints' (hint: "arg_constraints: Dict[<type>, <type>] = ...")
torch/jit/__init__.py:1: error: Cannot find implementation or library stub for module named 'torch._C'
torch/jit/__init__.py:15: error: Module 'torch.utils' has no attribute 'set_module'
torch/jit/__init__.py:70: error: Name 'Attribute' already defined on line 68
torch/jit/__init__.py:213: error: On Python 3 '{}'.format(b'abc') produces "b'abc'"; use !r if this is a desired behavior
torch/jit/__init__.py:215: error: On Python 3 '{}'.format(b'abc') produces "b'abc'"; use !r if this is a desired behavior
torch/jit/__init__.py:1524: error: Unsupported dynamic base class "with_metaclass"
torch/jit/__init__.py:1869: error: Name 'ScriptModule' already defined on line 1524
torch/jit/__init__.py:1998: error: Need type annotation for '_jit_caching_layer'
torch/jit/__init__.py:1999: error: Need type annotation for '_jit_function_overload_caching'
torch/distributions/relaxed_categorical.py:34: error: Incompatible types in assignment (expression has type "_Real", base class "Distribution" defined the type as "None")
torch/distributions/relaxed_categorical.py:108: error: Incompatible types in assignment (expression has type "_Simplex", base class "Distribution" defined the type as "None")
torch/distributions/relaxed_bernoulli.py:31: error: Incompatible types in assignment (expression has type "_Real", base class "Distribution" defined the type as "None")
torch/distributions/relaxed_bernoulli.py:114: error: Incompatible types in assignment (expression has type "_Interval", base class "Distribution" defined the type as "None")
torch/distributions/logistic_normal.py:31: error: Incompatible types in assignment (expression has type "_Simplex", base class "Distribution" defined the type as "None")
torch/distributions/log_normal.py:26: error: Incompatible types in assignment (expression has type "_GreaterThan", base class "Distribution" defined the type as "None")
torch/distributions/half_normal.py:27: error: Incompatible types in assignment (expression has type "_GreaterThan", base class "Distribution" defined the type as "None")
torch/distributions/half_cauchy.py:28: error: Incompatible types in assignment (expression has type "_GreaterThan", base class "Distribution" defined the type as "None")
torch/distributions/gumbel.py:28: error: Incompatible types in assignment (expression has type "_Real", base class "Distribution" defined the type as "None")
torch/nn/quantized/modules/conv.py:18: error: Module 'torch.nn.utils' has no attribute 'fuse_conv_bn_weights'
torch/nn/quantized/modules/conv.py:209: error: Name 'Optional' is not defined
torch/nn/quantized/modules/conv.py:209: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Optional")
torch/nn/quantized/modules/conv.py:214: error: Module has no attribute "ops"
torch/nn/quantized/modules/conv.py:321: error: Name 'Optional' is not defined
torch/nn/quantized/modules/conv.py:321: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Optional")
torch/nn/quantized/modules/conv.py:323: error: Module has no attribute "ops"
torch/nn/quantized/modules/conv.py:447: error: Name 'Optional' is not defined
torch/nn/quantized/modules/conv.py:447: note: Did you forget to import it from "typing"? (Suggestion: "from typing import Optional")
torch/nn/quantized/modules/conv.py:449: error: Module has no attribute "ops"
torch/nn/quantized/modules/conv.py:513: error: Name 'nn.modules.conv._ConvTransposeNd' is not defined
torch/nn/quantized/modules/conv.py:525: error: Name 'List' is not defined
torch/nn/quantized/modules/conv.py:525: note: Did you forget to import it from "typing"? (Suggestion: "from typing import List")
torch/nn/quantized/modules/conv.py:527: error: Name 'List' is not defined
torch/nn/quantized/modules/conv.py:527: note: Did you forget to import it from "typing"? (Suggestion: "from typing import List")
torch/nn/intrinsic/quantized/modules/conv_relu.py:8: error: Module 'torch.nn.utils' has no attribute 'fuse_conv_bn_weights'
torch/nn/intrinsic/quantized/modules/conv_relu.py:21: error: Incompatible types in assignment (expression has type "Type[ConvReLU2d]", base class "Conv2d" defined the type as "Type[Conv2d]")
torch/nn/intrinsic/quantized/modules/conv_relu.py:62: error: Incompatible types in assignment (expression has type "Type[ConvReLU3d]", base class "Conv3d" defined the type as "Type[Conv3d]")
torch/distributions/weibull.py:25: error: Incompatible types in assignment (expression has type "_GreaterThan", base class "Distribution" defined the type as "None")
torch/distributions/kl.py:35: error: Need type annotation for '_KL_MEMOIZE' (hint: "_KL_MEMOIZE: Dict[<type>, <type>] = ...")
torch/distributions/studentT.py:27: error: Incompatible types in assignment (expression has type "_Real", base class "Distribution" defined the type as "None")
torch/distributions/mixture_same_family.py:48: error: Need type annotation for 'arg_constraints' (hint: "arg_constraints: Dict[<type>, <type>] = ...")
torch/distributions/__init__.py:158: error: Name 'transforms' is not defined
torch/onnx/utils.py:21: error: Cannot find implementation or library stub for module named 'torch._C'
torch/distributed/rendezvous.py:4: error: Cannot find implementation or library stub for module named 'urlparse'
torch/distributed/rendezvous.py:4: error: Name 'urlparse' already defined (possibly by an import)
torch/distributed/rendezvous.py:4: error: Name 'urlunparse' already defined (possibly by an import)
torch/distributed/rendezvous.py:9: error: Module 'torch.distributed' has no attribute 'FileStore'
torch/distributed/rendezvous.py:9: error: Module 'torch.distributed' has no attribute 'TCPStore'
torch/distributed/rendezvous.py:65: error: On Python 3 '{}'.format(b'abc') produces "b'abc'"; use !r if this is a desired behavior
torch/distributed/distributed_c10d.py:11: error: Module 'torch.distributed' has no attribute 'AllreduceOptions'; maybe "ReduceOptions" or "AllreduceCoalescedOptions"?
torch/distributed/distributed_c10d.py:11: error: Module 'torch.distributed' has no attribute 'AllreduceCoalescedOptions'; maybe "AllreduceOptions"?
torch/distributed/distributed_c10d.py:11: error: Module 'torch.distributed' has no attribute 'AllToAllOptions'
torch/distributed/distributed_c10d.py:11: error: Module 'torch.distributed' has no attribute 'BroadcastOptions'
torch/distributed/distributed_c10d.py:11: error: Module 'torch.distributed' has no attribute 'GatherOptions'; maybe "ScatterOptions"?
torch/distributed/distributed_c10d.py:11: error: Module 'torch.distributed' has no attribute 'ReduceOptions'; maybe "AllreduceOptions", "ReduceScatterOptions", or "ReduceOp"?
torch/distributed/distributed_c10d.py:11: error: Module 'torch.distributed' has no attribute 'ReduceScatterOptions'; maybe "ScatterOptions" or "ReduceOptions"?
torch/distributed/distributed_c10d.py:11: error: Module 'torch.distributed' has no attribute 'ScatterOptions'; maybe "ReduceScatterOptions" or
Pull Request resolved: https://github.com/pytorch/pytorch/pull/36584
Reviewed By: seemethere, ailzhang
Differential Revision: D21155985
Pulled By: ezyang
fbshipit-source-id: f628d4293992576207167e7c417998fad15898d1
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/35615
Python 2 has reached end-of-life and is no longer supported by PyTorch.
Now we can clean up a lot of cruft that we put in place to support it.
These changes were all done manually, and I skipped anything that seemed
like it would take more than a few seconds, so I think it makes sense to
review it manually as well (though using side-by-side view and ignoring
whitespace change might be helpful).
Test Plan: CI
Differential Revision: D20842886
Pulled By: dreiss
fbshipit-source-id: 8cad4e87c45895e7ce3938a88e61157a79504aed
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/36741
Create child workspace that shares parent workspace's blobs. Register child workspace in registrar to enable switching into child workspace and feeding to child workspace alone.
Test Plan: numeric suite unit tests in stacked diff
Reviewed By: hx89
Differential Revision: D21055567
fbshipit-source-id: 374b12aef75a4c58452c271f8961ee156ce6c559
Summary: It was always skipped for last 1.5 years (since D10372230 was landed)
Test Plan: CI
Reviewed By: ailzhang
Differential Revision: D21036194
fbshipit-source-id: 9ace60b45a123a9372a88310b91f33a69ae8880c
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/36399
Added caffe2 python wrapper and unit test for the STORM C++ operator.
Test Plan:
All newly added unit tests passed using "buck test //caffe2/caffe2/python:optimizer_test -- TestStorm"
{F233644598}
Reviewed By: chocjy
Differential Revision: D18841013
fbshipit-source-id: f692bc18412839db140202ec9a971e556db0e54f
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/36225
Implemented the [STORM](https://arxiv.org/abs/1905.10018) optimizer operator for dense and sparse cases.
Test Plan:
All newly added unit tests passed using "buck test //caffe2/caffe2/python/operator_test:storm_test".
{F233643713}
Reviewed By: chocjy
Differential Revision: D18702897
fbshipit-source-id: d25eeb492aa2a03c69754d3f076a8239230b3bf4
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/36558
In the log, frequently see a large trunk of Using engine xx for rowWise Adagrad, but without information on which parameter is applied.
Test Plan: Should be covered by existing testing that use optimizer
Reviewed By: chocjy
Differential Revision: D20985176
fbshipit-source-id: 6eb4e19e5307db53fc89b38594a3f303f1492a1c
Summary:
Make the e2e FakeLowP python tests work with Glow lowering in OSS environment. Added a README.md as a guideline.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/36525
Reviewed By: hyuen
Differential Revision: D21004706
Pulled By: yinghai
fbshipit-source-id: d182152e4a1a3368640bd7872cb9ea4d4bff4b02
Summary:
We open sourced the FakeLowp ops as a reference implementation of fp16 ops. This PR makes it buildable.
```
USE_CUDA=0 USE_ROCM=0 USE_FAKELOWP=ON python setup.py install
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/36170
Test Plan:
Build Onnxifi library in Glow.
```
cp ${GLOW}/build/lib/Onnxifi/libonnxifi-glow.so ${MY_PATH}/ibonnxifi.so
LD_LIBRARY_PATH=${MY_PATH}/ibonnxifi.so python pytorch/caffe2/python/fakelowp/test_sls_nnpi_fp16.py
```
It doesn't run successfully right now because we need to open source the glow gflags and some other ops like `FbgemmPack`.
Reviewed By: houseroad
Differential Revision: D20980681
Pulled By: yinghai
fbshipit-source-id: 6dd31883a985850a77261bcc527029479bbc303f
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/36409
Pull Request resolved: https://github.com/pytorch/glow/pull/4409
Since glow OSS doesn't really ship with python, it's much easier to do it in pytorch. All the glow dependency can be done though LD_LIBRARY_PATH in OSS.
Test Plan:
```
buck test caffe2/caffe2/python/fakelowp:
```
Reviewed By: amylittleyang
Differential Revision: D20969308
fbshipit-source-id: 06a02d23f4972a92beb18e1d052e27d8724539d0
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/34527
Adding support for prune_delays and prune ratios in Adagrad optimizer.
Test Plan:
Tested via unit tests in masked_adagrad_optimizer_test. Added unit test for prune_delay versions of MaskedAdagrad
buck build caffe2/caffe2/fb/optimizers:masked_adagrad_optimizer_test; buck-out/gen/caffe2/caffe2/fb/optimizers/masked_adagrad_optimizer_test#binary.par
buck test caffe2/caffe2/fb/dper/layer_models/tests/split_1:sparse_nn_test -- 'test_pruning'
All Dper tests passed https://our.intern.facebook.com/intern/testinfra/testrun/7599824380741217
Reviewed By: chocjy
Differential Revision: D20313419
fbshipit-source-id: 5c2c8d4e0fc2ec538bcd6f145c6b87a2381f90f3
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/36267
This makes PythonOp throw the original python exception instead of wrapping it in a c10::Error type. This allows throwing exceptions from Python and preserving the type when they're caught again in Python. This is important for structured logging and handling non-retryable error types.
Test Plan: buck test caffe2/caffe2/python:python_op_test
Reviewed By: wenqicaofb
Differential Revision: D20928098
fbshipit-source-id: 001747f022c657b420f8450b84d64f4d57f6cdf6
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/35763
Adds inference function and test for ScatterAssign
Test Plan: Updated unit test
Reviewed By: yyetim, shunting1986
Differential Revision: D20501079
fbshipit-source-id: 7ec6ef0127a151250dd699c90c2b80c35cfb1fe4
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/35857
This fixes a lot of common ops for InferBlobShapesAndTypes as well as adds support for testing the inferred shapes and types of gradient ops.
Ops:
* Concat
* Split
* LeakyReLU
* Relu
* Prelu
* Gelu
* Elu
* Sinh, Tanh, Cosh
* Abs
* ... and a number of other simple element wise ops
Test Plan:
Added support to hypothesis test to check the shape and type of gradient ops.
Enabled it for all the ops I fixed the shape and type inference for.
buck test caffe2/caffe2/python/operator_test:
Reviewed By: pradeepd24
Differential Revision: D20806284
fbshipit-source-id: 77f796d9ff208e09e871bdbadf9a0a7c196b77f2
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/35555
Att. So that we can lower the SparseLengthsSum* part of SparseLengthsSum*Sparse. We update the tying policy between Gather and SparsLengthsWeightSum* so that we don't bother lowering a single Gather into the backend, which is inefficient to execute on card and creates bubbles between continuous lowering graphs.
Test Plan:
```
buck test glow/fb/test:test_onnxifinnpi
```
Reviewed By: ipiszy
Differential Revision: D20688525
fbshipit-source-id: cb8e38239057ff13a8d385ed09d0d019421de78b
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/35507
We want to split up the SparseLengthsSumSparse op into an indirection op and the SparseLengthsSum op so that we can lower the later part. The indirection part is a plain impl now.
Test Plan:
```
for i in `seq 10`; do buck test caffe2/caffe2/python/operator_test:lengths_reducer_fused_nbit_rowwise_ops_test -- test_sparse_lengths_sum_rowwise_sparse; done
```
Reviewed By: jspark1105
Differential Revision: D20683478
fbshipit-source-id: 509effe88719d20aa0c4783bbe0ce1f183ee473c
Summary:
## Motivation
This PR upgrades MKL-DNN from v0.20 to DNNL v1.2 and resolves https://github.com/pytorch/pytorch/issues/30300.
DNNL (Deep Neural Network Library) is the new brand of MKL-DNN, which improves performance, quality, and usability over the old version.
This PR focuses on the migration of all existing functionalities, including minor fixes, performance improvement and code clean up. It serves as the cornerstone of our future efforts to accommodate new features like OpenCL support, BF16 training, INT8 inference, etc. and to let the Pytorch community derive more benefits from the Intel Architecture.
<br>
## What's included?
Even DNNL has many breaking changes to the API, we managed to absorb most of them in ideep. This PR contains minimalist changes to the integration code in pytorch. Below is a summary of the changes:
<br>
**General:**
1. Replace op-level allocator with global-registered allocator
```
// before
ideep::sum::compute<AllocForMKLDNN>(scales, {x, y}, z);
// after
ideep::sum::compute(scales, {x, y}, z);
```
The allocator is now being registeted at `aten/src/ATen/native/mkldnn/IDeepRegistration.cpp`. Thereafter all tensors derived from the `cpu_engine` (by default) will use the c10 allocator.
```
RegisterEngineAllocator cpu_alloc(
ideep::engine::cpu_engine(),
[](size_t size) {
return c10::GetAllocator(c10::DeviceType::CPU)->raw_allocate(size);
},
[](void* p) {
c10::GetAllocator(c10::DeviceType::CPU)->raw_deallocate(p);
}
);
```
------
2. Simplify group convolution
We had such a scenario in convolution where ideep tensor shape mismatched aten tensor: when `groups > 1`, DNNL expects weights tensors to be 5-d with an extra group dimension, e.g. `goihw` instead of `oihw` in 2d conv case.
As shown below, a lot of extra checks came with this difference in shape before. Now we've completely hidden this difference in ideep and all tensors are going to align with pytorch's definition. So we could safely remove these checks from both aten and c2 integration code.
```
// aten/src/ATen/native/mkldnn/Conv.cpp
if (w.ndims() == x.ndims() + 1) {
AT_ASSERTM(
groups > 1,
"Only group _mkldnn_conv2d weights could have been reordered to 5d");
kernel_size[0] = w.get_dim(0) * w.get_dim(1);
std::copy_n(
w.get_dims().cbegin() + 2, x.ndims() - 1, kernel_size.begin() + 1);
} else {
std::copy_n(w.get_dims().cbegin(), x.ndims(), kernel_size.begin());
}
```
------
3. Enable DNNL built-in cache
Previously, we stored DNNL jitted kernels along with intermediate buffers inside ideep using an LRU cache. Now we are switching to the newly added DNNL built-in cache, and **no longer** caching buffers in order to reduce memory footprint.
This change will be mainly reflected in lower memory usage from memory profiling results. On the code side, we removed couple of lines of `op_key_` that depended on the ideep cache before.
------
4. Use 64-bit integer to denote dimensions
We changed the type of `ideep::dims` from `vector<int32_t>` to `vector<int64_t>`. This renders ideep dims no longer compatible with 32-bit dims used by caffe2. So we use something like `{stride_.begin(), stride_.end()}` to cast parameter `stride_` into a int64 vector.
<br>
**Misc changes in each commit:**
**Commit:** change build options
Some build options were slightly changed, mainly to avoid name collisions with other projects that include DNNL as a subproject. In addition, DNNL built-in cache is enabled by option `DNNL_ENABLE_PRIMITIVE_CACHE`.
Old | New
-- | --
WITH_EXAMPLE | MKLDNN_BUILD_EXAMPLES
WITH_TEST | MKLDNN_BUILD_TESTS
MKLDNN_THREADING | MKLDNN_CPU_RUNTIME
MKLDNN_USE_MKL | N/A (not use MKL anymore)
------
**Commit:** aten reintegration
- aten/src/ATen/native/mkldnn/BinaryOps.cpp
Implement binary ops using new operation `binary` provided by DNNL
- aten/src/ATen/native/mkldnn/Conv.cpp
Clean up group convolution checks
Simplify conv backward integration
- aten/src/ATen/native/mkldnn/MKLDNNConversions.cpp
Simplify prepacking convolution weights
- test/test_mkldnn.py
Fixed an issue in conv2d unit test: it didn't check conv results between mkldnn and aten implementation before. Instead, it compared the mkldnn with mkldnn as the default cpu path will also go into mkldnn. Now we use `torch.backends.mkldnn.flags` to fix this issue
- torch/utils/mkldnn.py
Prepack weight tensor on module `__init__` to achieve better performance significantly
------
**Commit:** caffe2 reintegration
- caffe2/ideep/ideep_utils.h
Clean up unused type definitions
- caffe2/ideep/operators/adam_op.cc & caffe2/ideep/operators/momentum_sgd_op.cc
Unify tensor initialization with `ideep::tensor::init`. Obsolete `ideep::tensor::reinit`
- caffe2/ideep/operators/conv_op.cc & caffe2/ideep/operators/quantization/int8_conv_op.cc
Clean up group convolution checks
Revamp convolution API
- caffe2/ideep/operators/conv_transpose_op.cc
Clean up group convolution checks
Clean up deconv workaround code
------
**Commit:** custom allocator
- Register c10 allocator as mentioned above
<br><br>
## Performance
We tested inference on some common models based on user scenarios, and most performance numbers are either better than or on par with DNNL 0.20.
ratio: new / old | Latency (batch=1 4T) | Throughput (batch=64 56T)
-- | -- | --
pytorch resnet18 | 121.4% | 99.7%
pytorch resnet50 | 123.1% | 106.9%
pytorch resnext101_32x8d | 116.3% | 100.1%
pytorch resnext50_32x4d | 141.9% | 104.4%
pytorch mobilenet_v2 | 163.0% | 105.8%
caffe2 alexnet | 303.0% | 99.2%
caffe2 googlenet-v3 | 101.1% | 99.2%
caffe2 inception-v1 | 102.2% | 101.7%
caffe2 mobilenet-v1 | 356.1% | 253.7%
caffe2 resnet101 | 100.4% | 99.8%
caffe2 resnet152 | 99.8% | 99.8%
caffe2 shufflenet | 141.1% | 69.0% †
caffe2 squeezenet | 98.5% | 99.2%
caffe2 vgg16 | 136.8% | 100.6%
caffe2 googlenet-v3 int8 | 100.0% | 100.7%
caffe2 mobilenet-v1 int8 | 779.2% | 943.0%
caffe2 resnet50 int8 | 99.5% | 95.5%
_Configuration:
Platform: Skylake 8180
Latency Test: 4 threads, warmup 30, iteration 500, batch size 1
Throughput Test: 56 threads, warmup 30, iteration 200, batch size 64_
† Shufflenet is one of the few models that require temp buffers during inference. The performance degradation is an expected issue since we no longer cache any buffer in the ideep. As for the solution, we suggest users opt for caching allocator like **jemalloc** as a drop-in replacement for system allocator in such heavy workloads.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/32422
Test Plan:
Perf results: https://our.intern.facebook.com/intern/fblearner/details/177790608?tab=Experiment%20Results
10% improvement for ResNext with avx512, neutral on avx2
More results: https://fb.quip.com/ob10AL0bCDXW#NNNACAUoHJP
Reviewed By: yinghai
Differential Revision: D20381325
Pulled By: dzhulgakov
fbshipit-source-id: 803b906fd89ed8b723c5fcab55039efe3e4bcb77
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/35430
This fixes and adds tests for several commonly used operators.
There's some formatting differences due to running clang-format on one of the files.
Test Plan: buck test //caffe2/caffe2/fb/operators:hypothesis_test //caffe2/caffe2/python/operator_test:utility_ops_test //caffe2/caffe2/python/operator_test:concat_split_op_test
Reviewed By: yyetim
Differential Revision: D20657405
fbshipit-source-id: 51d86d0834003b8ac8d6acb5149ae13d7bbfc6ab
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/35346
weight scale op doesn't have GPU impl. This is breaking OSS CI from D20506032. Making it cpu only
Test Plan: OSS CI
Reviewed By: ustctf
Differential Revision: D20637440
fbshipit-source-id: 9aa6cce63ce637ab7856788e5d02f527decb2a26
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/34394
# SWA operator
In this diff, we added a new operator `SWA` which will be used in `AdaGradOptimizer`.
The algorithm looks like:
{F230902995}
# Background
In our testings, we found that this operator could improve our models' reproducibility a lot. (KT: 0.86 -> .92)
So we hope to land this operator and in future, enable this by default in our Models.
Test Plan:
Local build `aml.dper3:30f068668cfb408fbb40141fb17129f2` and bento kernel.
- Local test: n215857
- f174600345
Reviewed By: chocjy
Differential Revision: D20165239
fbshipit-source-id: c03cdd048cb10b091e5f06323f4c0f3999f95d8a
Summary: Add transfer_learning_blob_name_mappings into layer_model_helper to support layer model transfer learning
Reviewed By: mraway
Differential Revision: D20286298
fbshipit-source-id: de3e029611d843f38d3f42ecd4148358f7e14a2b
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/34903
Reattempt of D20461609
Moving 2/4-bit SLS and row-wise 2/4-bit conversion operator to open source to be used by DLRM
Test Plan: CI
Reviewed By: jianyuh
Differential Revision: D20495304
fbshipit-source-id: 66a99677583f50fd40e29c514710c7b1a8cdbc29
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/34783
Moving 2/4-bit SLS and row-wise 2/4-bit conversion operator to open source to be used by DLRM
Test Plan: CI
Reviewed By: yinghai
Differential Revision: D20461609
fbshipit-source-id: b3ef73ff10f2433afe06ffa73fe1145282d9ec4c
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/34515
Once upon a time we thought this was necessary. In reality it is not, so
removing it.
For backcompat, our public interface (defined in `api/`) still has
typedefs to the old `script::` names.
There was only one collision: `Pass` as a `Stmt` and `Pass` as a graph
transform. I renamed one of them.
Test Plan: Imported from OSS
Differential Revision: D20353503
Pulled By: suo
fbshipit-source-id: 48bb911ce75120a8c9e0c6fb65262ef775dfba93
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/34318
Stop checking whether we have AMD GPU devices on the host, because we may be constructing a net on a machine without GPU, and run the net on another one with GPU
Reviewed By: ajauhri
Differential Revision: D20269562
fbshipit-source-id: 1f561086cacdcead3ce7c03c2d02c25336c8b11a
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/33977
Removing python2 from operator_test so we can retire python2 support for PyTorch.
Test Plan: waitforsandcastle
Reviewed By: seemethere
Differential Revision: D20129500
fbshipit-source-id: d4c82e4acfc795be9bec6a162c713e37ffb9f5ff
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/33431
Some elementwise operators don't have shape and type inference specified for the output tensor: `BitwiseOr`, `BitwiseAnd`, `BitwiseXor`, `Not`, `Sign`.
This change fixes this issue:
- For `Not` and `Sign` operators, the output has the same type and shape as the input, so `IdenticalTypeAndShapeOfInput` function is used to specify that.
- For bitwise operators created by `CAFFE2_SCHEMA_FOR_BINARY_BITWISE_OP` macro, the type and shape inference rules should be the same as for other binary element-wise operators, so `TensorInferenceFunction(ElementwiseOpShapeInference)` is used to specify that.
Also some tests were modified to ensure that the shape and type are inferred (`ensure_outputs_are_inferred` parameter)
Test Plan:
```
CAFFE2_ASSERT_SHAPEINFERENCE=1 buck test caffe2/caffe2/python/operator_test:elementwise_ops_test
CAFFE2_ASSERT_SHAPEINFERENCE=1 buck test caffe2/caffe2/python/operator_test:math_ops_test
```
Note that the tests have to be executed with `CAFFE2_ASSERT_SHAPEINFERENCE=1` in order to fail upon shape inference failure.
Reviewed By: idning
Differential Revision: D19880164
fbshipit-source-id: 5d7902e045d79e5669e5e98dfb13a39711294939
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/33426
Make 2/4/8-bit fused rowwise conversion operators more general to work for N-dim tensors
Test Plan: CI
Reviewed By: ellie-wen
Differential Revision: D19943136
fbshipit-source-id: 47008544dd7e1d11a346d34f35449e0fcc0e7ee0
Summary: in dper2, local net is hard-coded by whitelisting some layers. Add SparseFeatureGating related layers to local net explicitly.
Test Plan:
* workflow: f167812211
* QRT: fall back looks normal
{F228442018}
Differential Revision: D19852280
fbshipit-source-id: 6fecc3d745c3f742d029575a7b9fe320618f1863
Summary:
For both the Caffe2 and PyTorch backends, enable 3D convolutions through MIOpen.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/33067
Reviewed By: BIT-silence
Differential Revision: D19880495
Pulled By: bddppq
fbshipit-source-id: 8f6f970910654c1c5aa871b48a04c1054875691c
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/32271
Use the 2-stage EmbeddingSpMDM interface in D19425982 to reduce the overhead of code cache lookup and lock contention.
Fix an issue in sparse_lengths_sum_benchmarks generating empty indices when average length is small like 1.
Test Plan: CI
Reviewed By: dskhudia
Differential Revision: D19425987
fbshipit-source-id: d5c5f0d46e0072403901809c31d516fa0f4b9b31
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/32448
Using binary search to compute the value for the given quantile among the input tensors.
Test Plan: Newly added unittests;
Reviewed By: jspark1105
Differential Revision: D19487604
fbshipit-source-id: 0dc6627b78d1310ac35b3f1d53b89cc89a697ece
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/32475
As title
Test Plan: CI
Reviewed By: houseroad
Differential Revision: D19508778
fbshipit-source-id: fd9ad63607535980505d155f3e3c3b7c6b95daf7
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/32086
np.clip(1, num_indices // 2, 10) -> np.clip(num_indices // 2, 1, 10)
Also change batchsize -> num_rows to match with what the variable actually does
Test Plan: CI
Reviewed By: hx89
Differential Revision: D19361521
fbshipit-source-id: 9ce864c7d7da046dc606afa5207da677ccf80f52
Summary:
Per discussion with Fei Tian, we need to add a `scale_init_value` to scale down the output of normalization such as batch-norm and layer-norm.
Currently we have `sparse_normalization_options` to normalize embedding pooling output. By default, scale = 1.0, we found it's better to set scale from 0.025 to 0.1 https://fb.quip.com/MiKUAibEaYhH
Besides, I am removing the tags from normalizers because it makes more sense to calculate norm ops in distributed trainers, not ps.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/31983
Test Plan:
Testing LN and BN after sum-pooling --
baseline f160348514
LN: f160348609
BN: f160348710
{F226106518}
Layer norm after sum-pooling fwd_net https://fburl.com/sa4j207n
Layer norm after dot-prod fwd_net https://fburl.com/twggwyvb
## Unit Tests
Testing normalization after pooling
```
buck test caffe2/caffe2/fb/dper/layer_models/tests/split_1:sparse_nn_test_4 -- test_sparse_pooling_batch_normalization
buck test caffe2/caffe2/fb/dper/layer_models/tests/split_1:sparse_nn_test_4 -- test_dense_sparse_pooling_batch_normalization
buck test caffe2/caffe2/fb/dper/layer_models/tests/split_1:sparse_nn_test_4 -- test_sparse_pooling_layer_normalization
buck test caffe2/caffe2/fb/dper/layer_models/tests/split_1:sparse_nn_test_4 -- test_dense_sparse_pooling_layer_normalization
```
Testing normalization after dot-prod
```
buck test caffe2/caffe2/fb/dper/layer_models/tests/split_1:sparse_nn_test -- test_last_layer_use_batch_norm
buck test caffe2/caffe2/fb/dper/layer_models/tests/split_1:sparse_nn_test -- test_last_layer_use_layer_norm
```
Differential Revision: D19277618
Pulled By: SilunWang
fbshipit-source-id: ea323e33e3647ba55d2e808ef09d94ad7b45b934
