[ONNX] Support list remove for onnx export (#51373)

[BowenBao]

Stack from ghstack:

• [ONNX] Modifications in remove inplace ops passes to better handle binary inplace ops (#51318) #51572 [ONNX] Modifications in remove inplace ops passes to better handle binary inplace ops ([ONNX] Modifications in remove inplace ops passes to better handle binary inplace ops #51318)

• fix bug (#51222) #51527 fix bug ([ONNX] Fix bug in scatter_add symbolic #51222)

• [ONNX] Support list remove for onnx export (#51373) #51526 [ONNX] Support list remove for onnx export ([ONNX] Support list remove for onnx export #51373)

• [ONNX] Fix opset 11 ConstantChunk with negative dim (#51396) #51525 [ONNX] Fix opset 11 ConstantChunk with negative dim ([ONNX] Fix opset 11 ConstantChunk with negative dim #51396)

• [ONNX] Update unsafe_chunk() method to support new version 13 of Split operator. (#51415) #51524 [ONNX] Update unsafe_chunk() method to support new version 13 of Split operator. ([ONNX] Update unsafe_chunk() method to support new version 13 of Split operator. #51415)

• [ONNX] Enable Constant Folding for ONNX Opset 13 (#51096) #51523 [ONNX] Enable Constant Folding for ONNX Opset 13 ([ONNX] Enable Constant Folding for ONNX Opset 13 #51096)

• Update error message that displays when encountering an op unsupported for ONNX export. (#51387) #51522 Update error message that displays when encountering an op unsupported for ONNX export. (Update error message that displays when encountering an op unsupported for ONNX export. #51387)

• [ONNX] Fix graph sequence output from loop node (#51305) #51521 [ONNX] Fix graph sequence output from loop node ([ONNX] Fix graph sequence output from loop node #51305)

• [ONNX] Fix graph position to insert clone node for inplace op removal (#50123) #51520 [ONNX] Fix graph position to insert clone node for inplace op removal ([ONNX] Fix graph position to insert clone node for inplace op removal #50123)

• [ONNX] Add silu operator support for onnx (#51193) #51519 [ONNX] Add silu operator support for onnx ([ONNX] Add silu operator support for onnx #51193)

• [ONNX] Enable remaining failed tests in opset13 (#50806) #51518 [ONNX] Enable remaining failed tests in opset13 ([ONNX] Enable remaining failed tests in opset13 #50806)

• [ONNX] Export get/set attribute nodes (#50768) #51517 [ONNX] Export get/set attribute nodes ([ONNX] Export get/set attribute nodes  #50768)

• [ONNX] Improve error message for parse_arg in symbolic functions (#50512) #51516 [ONNX] Improve error message for parse_arg in symbolic functions ([ONNX] Improve error message for parse_arg in symbolic functions #50512)

• [ONNX] Fix bug in unfold symbolic (#50504) #51515 [ONNX] Fix bug in unfold symbolic ([ONNX] Fix bug in unfold symbolic  #50504)

• [ONNX] Update constant-folding of Gather op (#50554) #51514 [ONNX] Update constant-folding of Gather op ([ONNX] Update constant-folding of Gather op #50554)

• Support aten::Delete

• Refactor prepare_inplace_ops_for_onnx into one pass.

Differential Revision: D26203114

[facebook-github-bot]

💊 CI failures summary and remediations

As of commit 1d692cc (more details on the Dr. CI page):

---

• 2/2 failures possibly* introduced in this PR
  • 1/2 non-CircleCI failure(s)

---

🕵️ 1 new failure recognized by patterns

The following CI failures do not appear to be due to upstream breakages:

pytorch_linux_xenial_cuda10_2_cudnn7_py3_gcc7_test2 (1/1)

Step: "Run tests" (full log | diagnosis details | 🔁 rerun)

Feb 03 23:47:56 RuntimeError: CUDA error: an illegal memory access was encountered

Feb 03 23:47:56   File "test_optim.py", line 1994, in test_update_bn_dnn
Feb 03 23:47:56     self._test_update_bn(dnn.cuda(), dl_x, dl_xy, True)
Feb 03 23:47:56   File "/opt/conda/lib/python3.6/site-packages/torch/nn/modules/module.py", line 491, in cuda
Feb 03 23:47:56     return self._apply(lambda t: t.cuda(device))
Feb 03 23:47:56   File "/opt/conda/lib/python3.6/site-packages/torch/nn/modules/module.py", line 387, in _apply
Feb 03 23:47:56     module._apply(fn)
Feb 03 23:47:56   File "/opt/conda/lib/python3.6/site-packages/torch/nn/modules/module.py", line 409, in _apply
Feb 03 23:47:56     param_applied = fn(param)
Feb 03 23:47:56   File "/opt/conda/lib/python3.6/site-packages/torch/nn/modules/module.py", line 491, in <lambda>
Feb 03 23:47:56     return self._apply(lambda t: t.cuda(device))
Feb 03 23:47:56 RuntimeError: CUDA error: an illegal memory access was encountered
Feb 03 23:47:56 
Feb 03 23:47:56 ----------------------------------------------------------------------
Feb 03 23:47:56 Ran 104 tests in 38.216s
Feb 03 23:47:56 
Feb 03 23:47:56 FAILED (errors=13)
Feb 03 23:47:56 
Feb 03 23:47:56 Generating XML reports...
Feb 03 23:47:56 Generated XML report: test-reports/dist-gloo/TEST-TestLRScheduler-20210203234718.xml
Feb 03 23:47:56 Generated XML report: test-reports/dist-gloo/TEST-TestOptim-20210203234718.xml
Feb 03 23:47:56 Generated XML report: test-reports/dist-gloo/TEST-TestSWAUtils-20210203234718.xml

---

This comment was automatically generated by Dr. CI (expand for details).

Follow this link to opt-out of these comments for your Pull Requests.

Please report bugs/suggestions to the (internal) Dr. CI Users group.

[facebook-github-bot]

@SplitInfinity merged this pull request in 25b18bb.