[quant][fix] Fix quant type classification for float_qparam qconfig #48069

jerryzh168 · 2020-11-17T01:10:12Z

Stack from ghstack:

[quant][refactor] Move some util functions from torch/quantization/fx/utils.py to torch/quantization/utils.py #48107 [quant][refactor] Move some util functions from torch/quantization/fx/utils.py to torch/quantization/utils.py
[quant][fix][ez] Fix quant_type classification for fp16, fp16 #48073 [quant][fix][ez] Fix quant_type classification for fp16, fp16
[quant][fix] Fix quant type classification for float_qparam qconfig #48069 [quant][fix] Fix quant type classification for float_qparam qconfig
[quant][graphmode][fx] Keep linear op unchanged when qconfig is not supported #48067 [quant][graphmode][fx] Keep linear op unchanged when qconfig is not supported

Summary:
also renamed float_qparam_dynamic_qconfig to float_qparam_weight_only_qconfig
It's not used in user code yet so we only need to update the tests.

Test Plan:

Reviewers:

Subscribers:

Tasks:

Tags:

Differential Revision: D25010175

Summary: also renamed float_qparam_dynamic_qconfig to float_qparam_weight_only_qconfig It's not used in user code yet so we only need to update the tests. Test Plan: Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]

dr-ci · 2020-11-17T01:22:01Z

💊 CI failures summary and remediations

As of commit f35cdc6 (more details on the Dr. CI page):

3/3 failures introduced in this PR

🕵️ 3 new failures recognized by patterns

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

pytorch_linux_xenial_cuda9_2_cudnn7_py3_gcc5_4_build (1/3)

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

Nov 18 01:09:51 /var/lib/jenkins/workspace/torch/csrc/jit/mobile/import.h:16:32: error: unknown option after '#pragma GCC diagnostic' kind [-Werror=pragmas]

Nov 18 01:09:46 [ 75%] Building CXX object caffe2/CMakeFiles/torch_cpu.dir/__/torch/csrc/jit/codegen/fuser/cpu/fused_kernel.cpp.o 
Nov 18 01:09:46 [ 76%] Building CXX object caffe2/CMakeFiles/torch_cpu.dir/__/torch/csrc/jit/api/module_save.cpp.o 
Nov 18 01:09:47 [ 76%] Building CXX object caffe2/CMakeFiles/torch_cpu.dir/__/torch/csrc/utils/byte_order.cpp.o 
Nov 18 01:09:47 [ 76%] Building CXX object caffe2/CMakeFiles/torch_cpu.dir/__/torch/csrc/jit/serialization/import_legacy.cpp.o 
Nov 18 01:09:48 [ 76%] Building CXX object caffe2/CMakeFiles/torch_cpu.dir/__/torch/csrc/distributed/autograd/autograd.cpp.o 
Nov 18 01:09:49 [ 76%] Building CXX object caffe2/CMakeFiles/torch_cpu.dir/__/torch/csrc/distributed/autograd/utils.cpp.o 
Nov 18 01:09:49 [ 76%] Building CXX object caffe2/CMakeFiles/torch_cpu.dir/__/torch/csrc/distributed/autograd/context/container.cpp.o 
Nov 18 01:09:50 [ 76%] Building CXX object caffe2/CMakeFiles/torch_cpu.dir/__/torch/csrc/distributed/autograd/context/context.cpp.o 
Nov 18 01:09:51 [ 76%] Building CXX object caffe2/CMakeFiles/torch_cpu.dir/__/torch/csrc/distributed/autograd/engine/dist_engine.cpp.o 
Nov 18 01:09:51 In file included from /var/lib/jenkins/workspace/torch/csrc/jit/mobile/import.cpp:1:0: 
Nov 18 01:09:51 /var/lib/jenkins/workspace/torch/csrc/jit/mobile/import.h:16:32: error: unknown option after '#pragma GCC diagnostic' kind [-Werror=pragmas] 
Nov 18 01:09:51  #pragma GCC diagnostic ignored "-Wglobal-constructors" 
Nov 18 01:09:51                                 ^ 
Nov 18 01:09:51 cc1plus: all warnings being treated as errors 
Nov 18 01:09:51 make[2]: *** [caffe2/CMakeFiles/torch_cpu.dir/__/torch/csrc/jit/mobile/import.cpp.o] Error 1 
Nov 18 01:09:51 make[2]: *** Waiting for unfinished jobs.... 
Nov 18 01:09:51 caffe2/CMakeFiles/torch_cpu.dir/build.make:25454: recipe for target 'caffe2/CMakeFiles/torch_cpu.dir/__/torch/csrc/jit/mobile/import.cpp.o' failed 
Nov 18 01:09:52 make[1]: *** [caffe2/CMakeFiles/torch_cpu.dir/all] Error 2 
Nov 18 01:09:52 CMakeFiles/Makefile2:9923: recipe for target 'caffe2/CMakeFiles/torch_cpu.dir/all' failed 
Nov 18 01:09:52 Makefile:138: recipe for target 'all' failed 
Nov 18 01:09:52 make: *** [all] Error 2

binary_linux_libtorch_3_7m_cpu_gcc5_4_cxx11-abi_shared-with-deps_build (2/3)

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

RuntimeError: no cmake or cmake3 with version >= 3.5.0 found

++ [[ patchelf 0.12 == \p\a\t\c\h\e\l\f\ \0\.\9 ]] 
++ python setup.py clean 
Building wheel torch-1.8.0.dev20201118+cpu 
Traceback (most recent call last): 
  File "setup.py", line 284, in <module> 
    cmake = CMake() 
  File "/pytorch/tools/setup_helpers/cmake.py", line 98, in __init__ 
    self._cmake_command = CMake._get_cmake_command() 
  File "/pytorch/tools/setup_helpers/cmake.py", line 125, in _get_cmake_command 
    raise RuntimeError('no cmake or cmake3 with version >= 3.5.0 found') 
RuntimeError: no cmake or cmake3 with version >= 3.5.0 found

pytorch_linux_xenial_py3_6_gcc5_4_build (3/3)

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

Nov 18 01:04:42 /var/lib/jenkins/workspace/torch/csrc/jit/mobile/import.h:16:32: error: unknown option after '#pragma GCC diagnostic' kind [-Werror=pragmas]

Nov 18 01:04:37 [ 85%] Building CXX object caffe2/CMakeFiles/torch_cpu.dir/__/torch/csrc/utils/byte_order.cpp.o 
Nov 18 01:04:38 [ 85%] Building CXX object caffe2/CMakeFiles/torch_cpu.dir/__/torch/csrc/jit/serialization/import_legacy.cpp.o 
Nov 18 01:04:38 [ 85%] Building CXX object caffe2/CMakeFiles/torch_cpu.dir/__/torch/csrc/distributed/autograd/autograd.cpp.o 
Nov 18 01:04:39 [ 85%] Building CXX object caffe2/CMakeFiles/torch_cpu.dir/__/torch/csrc/distributed/autograd/utils.cpp.o 
Nov 18 01:04:39 [ 85%] Building CXX object caffe2/CMakeFiles/torch_cpu.dir/__/torch/csrc/distributed/autograd/context/container.cpp.o 
Nov 18 01:04:40 [ 85%] Building CXX object caffe2/CMakeFiles/torch_cpu.dir/__/torch/csrc/distributed/autograd/context/context.cpp.o 
Nov 18 01:04:40 [ 85%] Building CXX object caffe2/CMakeFiles/torch_cpu.dir/__/torch/csrc/distributed/autograd/engine/dist_engine.cpp.o 
Nov 18 01:04:41 [ 85%] Building CXX object caffe2/CMakeFiles/torch_cpu.dir/__/torch/csrc/distributed/autograd/functions/recvrpc_backward.cpp.o 
Nov 18 01:04:41 [ 85%] Building CXX object caffe2/CMakeFiles/torch_cpu.dir/__/torch/csrc/distributed/autograd/functions/sendrpc_backward.cpp.o 
Nov 18 01:04:42 In file included from /var/lib/jenkins/workspace/torch/csrc/jit/mobile/import.cpp:1:0: 
Nov 18 01:04:42 /var/lib/jenkins/workspace/torch/csrc/jit/mobile/import.h:16:32: error: unknown option after '#pragma GCC diagnostic' kind [-Werror=pragmas] 
Nov 18 01:04:42  #pragma GCC diagnostic ignored "-Wglobal-constructors" 
Nov 18 01:04:42                                 ^ 
Nov 18 01:04:42 cc1plus: all warnings being treated as errors 
Nov 18 01:04:42 caffe2/CMakeFiles/torch_cpu.dir/build.make:25406: recipe for target 'caffe2/CMakeFiles/torch_cpu.dir/__/torch/csrc/jit/mobile/import.cpp.o' failed 
Nov 18 01:04:42 make[2]: *** [caffe2/CMakeFiles/torch_cpu.dir/__/torch/csrc/jit/mobile/import.cpp.o] Error 1 
Nov 18 01:04:42 make[2]: *** Waiting for unfinished jobs.... 
Nov 18 01:04:42 make[1]: *** [caffe2/CMakeFiles/torch_cpu.dir/all] Error 2 
Nov 18 01:04:42 CMakeFiles/Makefile2:8620: recipe for target 'caffe2/CMakeFiles/torch_cpu.dir/all' failed 
Nov 18 01:04:42 Makefile:138: recipe for target 'all' failed 
Nov 18 01:04:42 make: *** [all] Error 2

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 or post in the (internal) Dr. CI Users group.

See how this bot performed.

This comment has been revised 42 times.

…m qconfig" Summary: also renamed float_qparam_dynamic_qconfig to float_qparam_weight_only_qconfig It's not used in user code yet so we only need to update the tests. Test Plan: Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D25010175](https://our.internmc.facebook.com/intern/diff/D25010175) [ghstack-poisoned]

Summary: also renamed float_qparam_dynamic_qconfig to float_qparam_weight_only_qconfig It's not used in user code yet so we only need to update the tests. Test Plan: Reviewers: Subscribers: Tasks: Tags: ghstack-source-id: 59ab7a6325ced695d826488a69babb92595dd6f3 Pull Request resolved: #48069

…m qconfig" Summary: also renamed float_qparam_dynamic_qconfig to float_qparam_weight_only_qconfig It's not used in user code yet so we only need to update the tests. Test Plan: Reviewers: Subscribers: Tasks: Tags: Differential Revision: [D25010175](https://our.internmc.facebook.com/intern/diff/D25010175) [ghstack-poisoned]

facebook-github-bot · 2020-11-19T03:13:49Z

This pull request has been merged in 576fa09.

jerryzh168 requested a review from apaszke as a code owner November 17, 2020 01:10

This was referenced Nov 17, 2020

[qunat][graphmode][fx] Embedding/EmbeddingBag works in static quant qconfig #48062

Closed

[quant][graphmode][fx] Keep linear op unchanged when qconfig is not supported #48067

Closed

facebook-github-bot added cla signed fx labels Nov 17, 2020

jerryzh168 mentioned this pull request Nov 17, 2020

[quant][fix][ez] Fix quant_type classification for fp16, fp16 #48073

Closed

jerryzh168 mentioned this pull request Nov 17, 2020

[quant][refactor] Move some util functions from torch/quantization/fx/utils.py to torch/quantization/utils.py #48107

Closed

supriyar approved these changes Nov 17, 2020

View reviewed changes

jerryzh168 added 5 commits November 17, 2020 10:16

facebook-github-bot closed this in 576fa09 Nov 19, 2020

facebook-github-bot added the Merged label Nov 19, 2020

facebook-github-bot deleted the gh/jerryzh168/507/head branch November 22, 2020 15:18

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[quant][fix] Fix quant type classification for float_qparam qconfig #48069

[quant][fix] Fix quant type classification for float_qparam qconfig #48069

jerryzh168 commented Nov 17, 2020 •

edited

dr-ci bot commented Nov 17, 2020 •

edited

facebook-github-bot commented Nov 19, 2020

[quant][fix] Fix quant type classification for float_qparam qconfig #48069

[quant][fix] Fix quant type classification for float_qparam qconfig #48069

Conversation

jerryzh168 commented Nov 17, 2020 • edited

dr-ci bot commented Nov 17, 2020 • edited

💊 CI failures summary and remediations

🕵️ 3 new failures recognized by patterns

pytorch_linux_xenial_cuda9_2_cudnn7_py3_gcc5_4_build (1/3)

binary_linux_libtorch_3_7m_cpu_gcc5_4_cxx11-abi_shared-with-deps_build (2/3)

pytorch_linux_xenial_py3_6_gcc5_4_build (3/3)

facebook-github-bot commented Nov 19, 2020

jerryzh168 commented Nov 17, 2020 •

edited

dr-ci bot commented Nov 17, 2020 •

edited