…7894eb

Summary:
Previous import was bff0b8835870c7df7762ef43498d000d2d8ffb52

Included changes:
- **[39dd0d4](onnx/onnx@39dd0d4)**: [build] Add ONNX_API for protos in all cases (pytorch#1407) <Orion Reblitz-Richardson>
- **[944db4f](onnx/onnx@944db4f)**: cmake (pytorch#1401) <zrphercule>
- **[8ccc8dd](onnx/onnx@8ccc8dd)**: Remove ONNXIFI_CHECK_RESULT from onnxRelease* functions (pytorch#1397) <Marat Dukhan>
- **[df14e74](onnx/onnx@df14e74)**: Change onnxifi test driver classname (pytorch#1396) <zrphercule>
- **[0c885cc](onnx/onnx@0c885cc)**: ONNXIFI cpp test driver (pytorch#1290) <zrphercule>
- **[a557848](onnx/onnx@a557848)**: Coverage Report Tools for Backend Scoreboard (pytorch#1301) <Akshay Chalana>
- **[31fd87f](onnx/onnx@31fd87f)**: fix AvgPool doc. add default value for count_include_pad (pytorch#1391) <Wenhao Hu>
- **[8ff08c2](onnx/onnx@8ff08c2)**: Do not export onnx symbols in the python extension (pytorch#1388) <bddppq>

Differential Revision: D9806635

fbshipit-source-id: 962e5dcb79f98a7e3a769b1ca9633e60c1735b48

…7894eb (#11622)

Summary:
Pull Request resolved: #11622

Previous import was bff0b8835870c7df7762ef43498d000d2d8ffb52

Included changes:
- **[39dd0d4](onnx/onnx@39dd0d4)**: [build] Add ONNX_API for protos in all cases (#1407) <Orion Reblitz-Richardson>
- **[944db4f](onnx/onnx@944db4f)**: cmake (#1401) <zrphercule>
- **[8ccc8dd](onnx/onnx@8ccc8dd)**: Remove ONNXIFI_CHECK_RESULT from onnxRelease* functions (#1397) <Marat Dukhan>
- **[df14e74](onnx/onnx@df14e74)**: Change onnxifi test driver classname (#1396) <zrphercule>
- **[0c885cc](onnx/onnx@0c885cc)**: ONNXIFI cpp test driver (#1290) <zrphercule>
- **[a557848](onnx/onnx@a557848)**: Coverage Report Tools for Backend Scoreboard (#1301) <Akshay Chalana>
- **[31fd87f](onnx/onnx@31fd87f)**: fix AvgPool doc. add default value for count_include_pad (#1391) <Wenhao Hu>
- **[8ff08c2](onnx/onnx@8ff08c2)**: Do not export onnx symbols in the python extension (#1388) <bddppq>

Reviewed By: orionr

Differential Revision: D9806635

fbshipit-source-id: f61c052b6bd14e0c80ace19c1a5f0ba659030c6f

* Have Kernel Inherit IrContainer (pytorch#1375)
* Kernel<-Fusion Step 1 - Convert ExprSort to StmtSort (pytorch#1376)
* Kernel<-Fusion Step 2 - Mutator refactor (pytorch#1377)
* Kernel<-Fusion Step 3 - Debug print for expr_eval and type promotion fix (pytorch#1379)
* Kernel<-Fusion Step 4 - Have kernel inherit Fusion (pytorch#1380)
* Kernel<-Fusion Step 5 - Move lowering passes into their own files (pytorch#1382)
* Kernel<-Fusion Step 6 - Remove kir::IrBuilder (pytorch#1383)
* Kernel<-Fusion Step 7 - Remove kir functions from ComputeAtMap (pytorch#1384)
* Kernel<-Fusion Step 8 - Clean up [lower/executor] utils (pytorch#1387)
* Kernel<-Fusion Step 9 - Remove TensorView::fuserTv (pytorch#1388)
* Kernel<-Fusion Step 10 - Remove lowerVal/lowerExpr (pytorch#1389)
* Kernel<-Fusion Step 11 - Finish cleaning up kir (pytorch#1390)

…orch#1388)

Some MIOpen RNN functions (lstm, rnn, gru) can't work with empty tensors and return error "MIOpen Error: Lengths must be > 0"
This PR disables MIOpen tor empty tensors and force to use native methods
The solution is based on condition of using CUDNN https://github.com/pytorch/pytorch/blob/3a52147cc59b240737602d3d046080bbf6f567f1/aten/src/ATen/native/TensorProperties.cpp#L91
It also fix [test_nn.py::TestNN::test_RNN_input_size_zero](https://github.com/pytorch/pytorch/blob/29fa6fbc4eda6c02ecdfd73b74a8702187c4fc44/test/test_nn.py#L4592) on ROCM

Pull Request resolved: pytorch#117672
Approved by: https://github.com/cpuhrsch