Change input scale to double type for conv params.

[kimishpatel]

Stack from ghstack:

• Some TODO fixes. #37829 Some TODO fixes.
• Perf optimization for conv and gemm kernels. #37626 Perf optimization for conv and gemm kernels.
• Added per channel separate test cases for fc and deconv tests. #37624 Added per channel separate test cases for fc and deconv tests.
• Changes to enable per channel support on dynamic linear. #37623 Changes to enable per channel support on dynamic linear.
• Enabled per channel quantized static linear/conv #37622 Enabled per channel quantized static linear/conv
• Added per channel kernels for depthwise conv. #37621 Added per channel kernels for depthwise conv.
• Changes to enable per channel requant. #37620 Changes to enable per channel requant.
• Enable per channel zero point. #37619 Enable per channel zero point.
• Interface changes to enable per channel quant. #37618 Interface changes to enable per channel quant.
• Change input scale to double type for conv params. #38346 Change input scale to double type for conv params.

Summary:
Given qtensor stores scale as double, this mismatch can cause use to
repack weights everytime in QNNPACK. Worse given that we release
original weights runtime can crash.

Test Plan:
pytest test/quantization/test_quantized_module.py::TestStaticQuantizedModule::test_conv2d_api

Reviewers:

Subscribers:

Tasks:

Tags:

Differential Revision: D21529384

[dr-ci]

💊 CI failures summary and remediations

As of commit 46a3f35 (more details on the Dr. CI page):

---

• 3/3 failures possibly* introduced in this PR
  • 2/3 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_test (1/1)

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

May 12 21:06:24 [E request_callback_impl.cpp:99] Received error while processing request type 15: size mismatch, m1: [3 x 3], m2: [6 x 6] at /var/lib/jenkins/workspace/aten/src/TH/generic/THTensorMath.cpp:41

May 12 21:06:20   test_context_cleanup_nested_rpc (__main__.DistAutogradTestWithSpawn) ... ok (0.813s) 
May 12 21:06:21   test_context_cleanup_no_tensors (__main__.DistAutogradTestWithSpawn) ... ok (0.813s) 
May 12 21:06:21   test_context_cleanup_tensor_no_grad (__main__.DistAutogradTestWithSpawn) ... ok (0.713s) 
May 12 21:06:22   test_context_cleanup_tensor_with_grad (__main__.DistAutogradTestWithSpawn) ... ok (0.813s) 
May 12 21:06:22   test_debug_info (__main__.DistAutogradTestWithSpawn) ... skip (0.003s) 
May 12 21:06:23   test_dist_autograd_profiling (__main__.DistAutogradTestWithSpawn) ... ok (0.914s) 
May 12 21:06:23   test_embedding_bag_with_no_grad_tensors (__main__.DistAutogradTestWithSpawn) ... skip (0.002s) 
May 12 21:06:24   test_error_in_context (__main__.DistAutogradTestWithSpawn) ... [E request_callback_impl.cpp:99] Received error while processing request type 15: size mismatch, m1: [3 x 3], m2: [6 x 6] at /var/lib/jenkins/workspace/aten/src/TH/generic/THTensorMath.cpp:41 
May 12 21:06:24 [E request_callback_impl.cpp:99] Received error while processing request type 15: size mismatch, m1: [3 x 3], m2: [6 x 6] at /var/lib/jenkins/workspace/aten/src/TH/generic/THTensorMath.cpp:41 
May 12 21:06:24 [E request_callback_impl.cpp:99] Received error while processing request type 15: size mismatch, m1: [3 x 3], m2: [6 x 6] at /var/lib/jenkins/workspace/aten/src/TH/generic/THTensorMath.cpp:41 
May 12 21:06:24 [E request_callback_impl.cpp:99] Received error while processing request type 15: size mismatch, m1: [3 x 3], m2: [6 x 6] at /var/lib/jenkins/workspace/aten/src/TH/generic/THTensorMath.cpp:41 
May 12 21:06:24 ok (0.713s) 
May 12 21:06:25   test_grad_copy_sparse_indices_extra_ref (__main__.DistAutogradTestWithSpawn) ... [W pybind_utils.h:798] Warning: Using sparse tensors in TorchScript is experimental. Many optimization pathways have not been thoroughly tested with sparse tensors. Please include the fact that the network is running sparse tensors in any bug reports submitted. (function operator()) 
May 12 21:06:25 [W pybind_utils.h:798] Warning: Using sparse tensors in TorchScript is experimental. Many optimization pathways have not been thoroughly tested with sparse tensors. Please include the fact that the network is running sparse tensors in any bug reports submitted. (function operator()) 
May 12 21:06:25 [W pybind_utils.h:798] Warning: Using sparse tensors in TorchScript is experimental. Many optimization pathways have not been thoroughly tested with sparse tensors. Please include the fact that the network is running sparse tensors in any bug reports submitted. (function operator()) 
May 12 21:06:25 [W pybind_utils.h:798] Warning: Using sparse tensors in TorchScript is experimental. Many optimization pathways have not been thoroughly tested with sparse tensors. Please include the fact that the network is running sparse tensors in any bug reports submitted. (function operator()) 
May 12 21:06:25 /opt/conda/lib/python3.6/site-packages/torch/nn/functional.py:1909: UserWarning: Argument order of nn.functional.embedding_bag was changed. Usage `embedding_bag(weight, input, ...)` is deprecated, and should now be `embedding_bag(input, weight, ...)`. 
May 12 21:06:25   warnings.warn("Argument order of nn.functional.embedding_bag was changed. " 
May 12 21:06:25 /opt/conda/lib/python3.6/site-packages/torch/nn/functional.py:1909: UserWarning: Argument order of nn.functional.embedding_bag was changed. Usage `embedding_bag(weight, input, ...)` is deprecated, and should now be `embedding_bag(input, weight, ...)`. 
May 12 21:06:25   warnings.warn("Argument order of nn.functional.embedding_bag was changed. " 
May 12 21:06:25 /opt/conda/lib/python3.6/site-packages/torch/nn/functional.py:1909: UserWarning: Argument order of nn.functional.embedding_bag was changed. Usage `embedding_bag(weight, input, ...)` is deprecated, and should now be `embedding_bag(input, weight, ...)`. 

---

ci.pytorch.org: 2 failed

• Failed: pr/py3.6-clang7-rocmdeb-ubuntu16.04
• Failed: pr/caffe2-py3.6-devtoolset7-rocmrpm-centos7-test

---

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 on the GitHub issue tracker.

See how this bot performed.

This comment has been revised 2 times.

[facebook-github-bot]

This pull request has been merged in 2c88141.