neo-ai / tvm Goto Github PK

This project forked from apache/tvm

Open deep learning compiler stack for cpu, gpu and specialized accelerators

License: Apache License 2.0

CMake 0.77% Makefile 0.22% Java 0.67% Shell 0.76% C++ 38.85% Python 54.95% Objective-C 0.08% Objective-C++ 0.29% Rust 1.29% Go 0.37% C 1.18% JavaScript 0.05% HTML 0.01% RenderScript 0.01% Cuda 0.06% TypeScript 0.32% Batchfile 0.01% Cython 0.10%

compiler deep-learning performance machine-learning

tvm's Introduction

Open Deep Learning Compiler Stack

Documentation | Contributors | Community | Release Notes

Apache TVM is a compiler stack for deep learning systems. It is designed to close the gap between the productivity-focused deep learning frameworks, and the performance- and efficiency-focused hardware backends. TVM works with deep learning frameworks to provide end to end compilation to different backends.

Neo-AI/TVM is a downstream branch of TVM that includes vendor- and product-specific features on top of the upstream codebase.

Branches

dev - This is the development branch with most update to date source code.

License

TVM is licensed under the Apache-2.0 license.

Getting Started

Check out the TVM Documentation site for installation instructions, tutorials, examples, and more. The Getting Started with TVM tutorial is a great place to start.

Contribute to TVM

TVM adopts apache committer model, we aim to create an open source project that is maintained and owned by the community. Check out the Contributor Guide.

Acknowledgement

We learned a lot from the following projects when building TVM.

Halide: Part of TVM's TIR and arithmetic simplification module originates from Halide. We also learned and adapted some part of lowering pipeline from Halide.
Loopy: use of integer set analysis and its loop transformation primitives.
Theano: the design inspiration of symbolic scan operator for recurrence.

tvm's People

Contributors

Stargazers

Watchers

tvm's Issues

task_python_vta.py failed

https://neo-ai-ci.amazon-ml.com/blue/organizations/jenkins/tvm/detail/PR-7/7/pipeline/36

+ docker/bash.sh tvmai/ci-cpu ./tests/scripts/task_python_vta.sh

WORKSPACE: /home/ubuntu/workspace/tvm/build-cpu

DOCKER CONTAINER NAME: tvmai/ci-cpu



Running './tests/scripts/task_python_vta.sh' inside tvmai/ci-cpu...

docker

mesg: ttyname failed: Inappropriate ioctl for device

Adding group `ubuntu' (GID 1000) ...

Done.

cd python; python setup.py build_ext --inplace

/usr/local/lib/python2.7/dist-packages/Cython/Compiler/Main.py:367: FutureWarning: Cython directive 'language_level' not set, using 2 for now (Py2). This will change in a later release! File: /workspace/python/tvm/_ffi/_cython/core.pyx

  tree = Parsing.p_module(s, pxd, full_module_name)

Compiling tvm/_ffi/_cython/core.pyx because it changed.

[1/1] Cythonizing tvm/_ffi/_cython/core.pyx

/usr/lib/python2.7/dist-packages/setuptools/dist.py:285: UserWarning: Normalizing '0.5.dev' to '0.5.dev0'

  normalized_version,

running build_ext

building 'tvm._ffi._cy2.core' extension

creating build

creating build/temp.linux-x86_64-2.7

creating build/temp.linux-x86_64-2.7/tvm

creating build/temp.linux-x86_64-2.7/tvm/_ffi

creating build/temp.linux-x86_64-2.7/tvm/_ffi/_cython

x86_64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -fno-strict-aliasing -Wdate-time -D_FORTIFY_SOURCE=2 -g -fstack-protector-strong -Wformat -Werror=format-security -fPIC -I../include/ -I../3rdparty/dmlc-core/include -I../3rdparty/dlpack/include -I/usr/include/python2.7 -c tvm/_ffi/_cython/core.cpp -o build/temp.linux-x86_64-2.7/tvm/_ffi/_cython/core.o

cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++

creating build/lib.linux-x86_64-2.7

creating build/lib.linux-x86_64-2.7/tvm

creating build/lib.linux-x86_64-2.7/tvm/_ffi

creating build/lib.linux-x86_64-2.7/tvm/_ffi/_cy2

c++ -pthread -shared -Wl,-O1 -Wl,-Bsymbolic-functions -Wl,-Bsymbolic-functions -Wl,-z,relro -fno-strict-aliasing -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -Wdate-time -D_FORTIFY_SOURCE=2 -g -fstack-protector-strong -Wformat -Werror=format-security -Wl,-Bsymbolic-functions -Wl,-z,relro -Wdate-time -D_FORTIFY_SOURCE=2 -g -fstack-protector-strong -Wformat -Werror=format-security build/temp.linux-x86_64-2.7/tvm/_ffi/_cython/core.o -o build/lib.linux-x86_64-2.7/tvm/_ffi/_cy2/core.so

copying build/lib.linux-x86_64-2.7/tvm/_ffi/_cy2/core.so -> tvm/_ffi/_cy2

cd python; python3 setup.py build_ext --inplace

running build_ext

building 'tvm._ffi._cy3.core' extension

creating build/temp.linux-x86_64-3.6

creating build/temp.linux-x86_64-3.6/tvm

creating build/temp.linux-x86_64-3.6/tvm/_ffi

creating build/temp.linux-x86_64-3.6/tvm/_ffi/_cython

x86_64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I../include/ -I../3rdparty/dmlc-core/include -I../3rdparty/dlpack/include -I/usr/include/python3.6m -c tvm/_ffi/_cython/core.cpp -o build/temp.linux-x86_64-3.6/tvm/_ffi/_cython/core.o

cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++

creating build/lib.linux-x86_64-3.6

creating build/lib.linux-x86_64-3.6/tvm

creating build/lib.linux-x86_64-3.6/tvm/_ffi

creating build/lib.linux-x86_64-3.6/tvm/_ffi/_cy3

x86_64-linux-gnu-g++ -pthread -shared -Wl,-O1 -Wl,-Bsymbolic-functions -Wl,-Bsymbolic-functions -Wl,-z,relro -Wl,-Bsymbolic-functions -Wl,-z,relro -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 build/temp.linux-x86_64-3.6/tvm/_ffi/_cython/core.o -o build/lib.linux-x86_64-3.6/tvm/_ffi/_cy3/core.cpython-36m-x86_64-linux-gnu.so

copying build/lib.linux-x86_64-3.6/tvm/_ffi/_cy3/core.cpython-36m-x86_64-linux-gnu.so -> tvm/_ffi/_cy3

/usr/local/lib/python3.6/dist-packages/setuptools/dist.py:398: UserWarning: Normalizing '0.5.dev' to '0.5.dev0'

  normalized_version,

Running unittest...

Failure: NameError (name 'avg_pool2d_alter_layout' is not defined) ... ERROR

Failure: ImportError (cannot import name cpp) ... ERROR



======================================================================

ERROR: Failure: NameError (name 'avg_pool2d_alter_layout' is not defined)

----------------------------------------------------------------------

Traceback (most recent call last):

  File "/usr/local/lib/python2.7/dist-packages/nose/loader.py", line 418, in loadTestsFromName

    addr.filename, addr.module)

  File "/usr/local/lib/python2.7/dist-packages/nose/importer.py", line 47, in importFromPath

    return self.importFromDir(dir_path, fqname)

  File "/usr/local/lib/python2.7/dist-packages/nose/importer.py", line 94, in importFromDir

    mod = load_module(part_fqname, fh, filename, desc)

  File "/workspace/vta/tests/python/unittest/test_environment.py", line 1, in <module>

    import vta

  File "/workspace/vta/python/vta/__init__.py", line 20, in <module>

    from . import top

  File "/workspace/vta/python/vta/top/__init__.py", line 3, in <module>

    from .vta_conv2d import packed_conv2d, schedule_packed_conv2d

  File "/workspace/vta/python/vta/top/vta_conv2d.py", line 8, in <module>

    import topi

  File "/workspace/topi/python/topi/__init__.py", line 26, in <module>

    from . import cuda

  File "/workspace/topi/python/topi/cuda/__init__.py", line 14, in <module>

    from .pooling import schedule_pool, schedule_global_pool

  File "/workspace/topi/python/topi/cuda/pooling.py", line 136, in <module>

    @avg_pool2d_alter_layout.register(["cuda"])

NameError: name 'avg_pool2d_alter_layout' is not defined



======================================================================

ERROR: Failure: ImportError (cannot import name cpp)

----------------------------------------------------------------------

Traceback (most recent call last):

  File "/usr/local/lib/python2.7/dist-packages/nose/loader.py", line 418, in loadTestsFromName

    addr.filename, addr.module)

  File "/usr/local/lib/python2.7/dist-packages/nose/importer.py", line 47, in importFromPath

    return self.importFromDir(dir_path, fqname)

  File "/usr/local/lib/python2.7/dist-packages/nose/importer.py", line 94, in importFromDir

    mod = load_module(part_fqname, fh, filename, desc)

  File "/workspace/vta/tests/python/unittest/test_vta_insn.py", line 4, in <module>

    import topi

  File "/workspace/topi/python/topi/__init__.py", line 16, in <module>

    from . import cpp

ImportError: cannot import name cpp



----------------------------------------------------------------------

Ran 2 tests in 0.237s



FAILED (errors=2)

Terminated

script returned exit code 255

Support ONNX Operators: ScatterND and Range of YoloV5

Hi,

I use YoloV5 to ONNX. Then want to use the SageMaker Neo Compile ONNX.
I got the following error messages:

ClientError: OperatorNotImplemented:('The following operators are not supported for frontend ONNX: Range, ScatterND")

You can repeat the error via this notebook link

Flaky CI in task_python_docs.sh

Thanks for participating in the TVM community! We use https://discuss.tvm.ai for any general usage questions and discussions. The issue tracker is used for actionable items such as feature proposals discussion, roadmaps, and bug tracking. You are always welcomed to post on the forum first :)

Issues that are inactive for a period of time may get closed. We adopt this policy so that we won't lose track of actionable issues that may fall at the bottom of the pile. Feel free to reopen a new one if you feel there is an additional problem that needs attention when an old one gets closed.

For bug reports, to help the developer act on the issues, please include a description of your environment, preferably a minimum script to reproduce the problem.

For feature proposals, list clear, small actionable items so we can track the progress of the change.

When attempting to merge #88, noticed some flaky CI issues. The CI would hang when generating docs for tutorial from_caffe2.py. When running the test script locally in the CI container (v0.56 for neoai/ci-gpu), I saw that it would pass but also have some other errors occasionally which could mean a sphinx issue. It is likely that updating to v0.60 in the recent merge could solve the issues.

avg_pool2d count_include_pad flag not set correctly in TensorRT wrapper

In tensorrt_executor.cc, the AddPooling function does not call nvinfer1::IPoolingLayer::setAverageCountExcludesPadding, which is necessary to set count_include_pad in avg_pool2d. Thus, AverageCountExcludesPadding is true by default. For models such as Inception V3 which using average pooling (for multiple frameworks, including MXNet and PyTorch), AverageCountExcludesPadding should be set to false via nvinfer1::IPoolingLayer::setAverageCountExcludesPadding since count_include_pad is set to true for Inception V3. The result is that Inception V3 produces incorrect outputs for multiple frameworks. @reminisce

Fix Sphinx documentation generation failure

There is failure in sphinx documentation generation, I skipped it for now(21c0b8e).

mlas submodule does not exist

The submodule mlas repo in .gitmodules does not exist.

ANTLR cmake build failure

https://neo-ai-ci.amazon-ml.com/blue/organizations/jenkins/tvm/detail/PR-7/3/pipeline/36/

+ docker/bash.sh tvmai/ci-cpu ./tests/scripts/task_build.sh build -j2

WORKSPACE: /home/ubuntu/workspace/tvm/build-cpu

DOCKER CONTAINER NAME: tvmai/ci-cpu



Running './tests/scripts/task_build.sh build -j2' inside tvmai/ci-cpu...

docker

mesg: ttyname failed: Inappropriate ioctl for device

Adding group `ubuntu' (GID 1000) ...

Done.

-- The C compiler identification is GNU 5.4.0

-- The CXX compiler identification is GNU 5.4.0

-- Check for working C compiler: /usr/bin/cc

-- Check for working C compiler: /usr/bin/cc -- works

-- Detecting C compiler ABI info

-- Detecting C compiler ABI info - done

-- Detecting C compile features

-- Detecting C compile features - done

-- Check for working CXX compiler: /usr/bin/c++

-- Check for working CXX compiler: /usr/bin/c++ -- works

-- Detecting CXX compiler ABI info

-- Detecting CXX compiler ABI info - done

-- Detecting CXX compile features

-- Detecting CXX compile features - done

-- Performing Test SUPPORT_CXX11

-- Performing Test SUPPORT_CXX11 - Success

-- Build with RPC support...

-- Build with Graph runtime support...

-- Build with Graph runtime debug support...

-- Build VTA runtime with target: sim

-- Use llvm-config=llvm-config-4.0

-- /usr/lib/llvm-4.0/include

-- Found LLVM_INCLUDE_DIRS=/usr/lib/llvm-4.0/include

-- Found LLVM_DEFINITIONS= -DNDEBUG -D_GNU_SOURCE -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS

-- Found TVM_LLVM_VERSION=40

-- Build with LLVM 

-- Set TVM_LLVM_VERSION=40

CMake Error at cmake/modules/ANTLR.cmake:12 (list):

  list GET given empty list

Call Stack (most recent call first):

  CMakeLists.txt:190 (include)





-- Build with contrib.sort

-- Build with contrib.hybriddump

-- Configuring incomplete, errors occurred!

See also "/workspace/build/CMakeFiles/CMakeOutput.log".

See also "/workspace/build/CMakeFiles/CMakeError.log".

CI build failure

https://neo-ai-ci.amazon-ml.com/blue/organizations/jenkins/tvm/detail/PR-3/12/pipeline

+ docker/bash.sh tvmai/ci-i386 ./tests/scripts/task_build.sh build -j2

WORKSPACE: /home/ubuntu/workspace/tvm/build-i386

DOCKER CONTAINER NAME: tvmai/ci-i386



Running './tests/scripts/task_build.sh build -j2' inside tvmai/ci-i386...

docker

mesg: ttyname failed: Inappropriate ioctl for device

Adding group `ubuntu' (GID 1000) ...

Done.

-- The C compiler identification is GNU 5.4.0

-- The CXX compiler identification is GNU 5.4.0

-- Check for working C compiler: /usr/bin/cc

-- Check for working C compiler: /usr/bin/cc -- works

-- Detecting C compiler ABI info

-- Detecting C compiler ABI info - done

-- Detecting C compile features

-- Detecting C compile features - done

-- Check for working CXX compiler: /usr/bin/c++

-- Check for working CXX compiler: /usr/bin/c++ -- works

-- Detecting CXX compiler ABI info

-- Detecting CXX compiler ABI info - done

-- Detecting CXX compile features

-- Detecting CXX compile features - done

CMake Error at CMakeLists.txt:45 (getFromList):

  Unknown CMake command "getFromList".





-- Configuring incomplete, errors occurred!

See also "/workspace/build/CMakeFiles/CMakeOutput.log".

pt vgg flaky issue on ci gpu instance

We encountered this issue in pr #63. File this issue for tracking.

cc: @alexwong

ONNX Upsample model compile error

Hi!

I have a compilation problem with sagemaker neo. May I report it here?
When I try to convert an onnx graph like the one below, I get an error and cannot convert it.

graph(%input : Float(1:519168, 3:173056, 416:416, 416:1),
%basenet.slice1.0.weight : Float(64:27, 3:9, 3:3, 3:1),
%basenet.slice1.0.bias : Float(64:1),
%basenet.slice1.1.weight : Float(64:1),
%basenet.slice1.1.bias : Float(64:1),
%basenet.slice1.1.running_mean : Float(64:1),
%basenet.slice1.1.running_var : Float(64:1),
%basenet.slice1.3.weight : Float(64:576, 64:9, 3:3, 3:1),
%basenet.slice1.3.bias : Float(64:1),
%basenet.slice1.4.weight : Float(64:1),
%basenet.slice1.4.bias : Float(64:1),
%basenet.slice1.4.running_mean : Float(64:1),
%basenet.slice1.4.running_var : Float(64:1),
%basenet.slice1.7.weight : Float(128:576, 64:9, 3:3, 3:1),
%basenet.slice1.7.bias : Float(128:1),
%basenet.slice1.8.weight : Float(128:1),
%basenet.slice1.8.bias : Float(128:1),
%basenet.slice1.8.running_mean : Float(128:1),
%basenet.slice1.8.running_var : Float(128:1),
%basenet.slice1.10.weight : Float(128:1152, 128:9, 3:3, 3:1),
%basenet.slice1.10.bias : Float(128:1),
%basenet.slice1.11.weight : Float(128:1),
%basenet.slice1.11.bias : Float(128:1),
%basenet.slice1.11.running_mean : Float(128:1),
%basenet.slice1.11.running_var : Float(128:1),
%basenet.slice2.14.weight : Float(256:1152, 128:9, 3:3, 3:1),
%basenet.slice2.14.bias : Float(256:1),
%basenet.slice2.15.weight : Float(256:1),
%basenet.slice2.15.bias : Float(256:1),
%basenet.slice2.15.running_mean : Float(256:1),
%basenet.slice2.15.running_var : Float(256:1),
%basenet.slice2.17.weight : Float(256:2304, 256:9, 3:3, 3:1),
%basenet.slice2.17.bias : Float(256:1),
%basenet.slice2.18.weight : Float(256:1),
%basenet.slice2.18.bias : Float(256:1),
%basenet.slice2.18.running_mean : Float(256:1),
%basenet.slice2.18.running_var : Float(256:1),
%basenet.slice3.20.weight : Float(256:2304, 256:9, 3:3, 3:1),
%basenet.slice3.20.bias : Float(256:1),
%basenet.slice3.21.weight : Float(256:1),
%basenet.slice3.21.bias : Float(256:1),
%basenet.slice3.21.running_mean : Float(256:1),
%basenet.slice3.21.running_var : Float(256:1),
%basenet.slice3.24.weight : Float(512:2304, 256:9, 3:3, 3:1),
%basenet.slice3.24.bias : Float(512:1),
%basenet.slice3.25.weight : Float(512:1),
%basenet.slice3.25.bias : Float(512:1),
%basenet.slice3.25.running_mean : Float(512:1),
%basenet.slice3.25.running_var : Float(512:1),
%basenet.slice3.27.weight : Float(512:4608, 512:9, 3:3, 3:1),
%basenet.slice3.27.bias : Float(512:1),
%basenet.slice3.28.weight : Float(512:1),
%basenet.slice3.28.bias : Float(512:1),
%basenet.slice3.28.running_mean : Float(512:1),
%basenet.slice3.28.running_var : Float(512:1),
%basenet.slice4.30.weight : Float(512:4608, 512:9, 3:3, 3:1),
%basenet.slice4.30.bias : Float(512:1),
%basenet.slice4.31.weight : Float(512:1),
%basenet.slice4.31.bias : Float(512:1),
%basenet.slice4.31.running_mean : Float(512:1),
%basenet.slice4.31.running_var : Float(512:1),
%basenet.slice4.34.weight : Float(512:4608, 512:9, 3:3, 3:1),
%basenet.slice4.34.bias : Float(512:1),
%basenet.slice4.35.weight : Float(512:1),
%basenet.slice4.35.bias : Float(512:1),
%basenet.slice4.35.running_mean : Float(512:1),
%basenet.slice4.35.running_var : Float(512:1),
%basenet.slice4.37.weight : Float(512:4608, 512:9, 3:3, 3:1),
%basenet.slice4.37.bias : Float(512:1),
%basenet.slice4.38.weight : Float(512:1),
%basenet.slice4.38.bias : Float(512:1),
%basenet.slice4.38.running_mean : Float(512:1),
%basenet.slice4.38.running_var : Float(512:1),
%basenet.slice5.1.weight : Float(1024:4608, 512:9, 3:3, 3:1),
%basenet.slice5.1.bias : Float(1024:1),
%basenet.slice5.2.weight : Float(1024:1024, 1024:1, 1:1, 1:1),
%basenet.slice5.2.bias : Float(1024:1),
%upconv1.conv.0.weight : Float(512:1536, 1536:1, 1:1, 1:1),
%upconv1.conv.0.bias : Float(512:1),
%upconv1.conv.1.weight : Float(512:1),
%upconv1.conv.1.bias : Float(512:1),
%upconv1.conv.1.running_mean : Float(512:1),
%upconv1.conv.1.running_var : Float(512:1),
%upconv1.conv.3.weight : Float(256:4608, 512:9, 3:3, 3:1),
%upconv1.conv.3.bias : Float(256:1),
%upconv1.conv.4.weight : Float(256:1),
%upconv1.conv.4.bias : Float(256:1),
%upconv1.conv.4.running_mean : Float(256:1),
%upconv1.conv.4.running_var : Float(256:1)):
%103 : Float(1:11075584, 64:173056, 416:416, 416:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%input, %basenet.slice1.0.weight, %basenet.slice1.0.bias) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/modules/conv.py:416:0
%104 : Float(1:11075584, 64:173056, 416:416, 416:1) = onnx::BatchNormalization[epsilon=1.0000000000000001e-05, momentum=0.90000000000000002](%103, %basenet.slice1.1.weight, %basenet.slice1.1.bias, %basenet.slice1.1.running_mean, %basenet.slice1.1.running_var) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:2016:0
%105 : Float(1:11075584, 64:173056, 416:416, 416:1) = onnx::Relu(%104) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:1117:0
%106 : Float(1:11075584, 64:173056, 416:416, 416:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%105, %basenet.slice1.3.weight, %basenet.slice1.3.bias) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/modules/conv.py:416:0
%107 : Float(1:11075584, 64:173056, 416:416, 416:1) = onnx::BatchNormalization[epsilon=1.0000000000000001e-05, momentum=0.90000000000000002](%106, %basenet.slice1.4.weight, %basenet.slice1.4.bias, %basenet.slice1.4.running_mean, %basenet.slice1.4.running_var) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:2016:0
%108 : Float(1:11075584, 64:173056, 416:416, 416:1) = onnx::Relu(%107) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:1117:0
%109 : Float(1:2768896, 64:43264, 208:208, 208:1) = onnx::MaxPoolkernel_shape=[2, 2], pads=[0, 0, 0, 0], strides=[2, 2] # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:576:0
%110 : Float(1:5537792, 128:43264, 208:208, 208:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%109, %basenet.slice1.7.weight, %basenet.slice1.7.bias) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/modules/conv.py:416:0
%111 : Float(1:5537792, 128:43264, 208:208, 208:1) = onnx::BatchNormalization[epsilon=1.0000000000000001e-05, momentum=0.90000000000000002](%110, %basenet.slice1.8.weight, %basenet.slice1.8.bias, %basenet.slice1.8.running_mean, %basenet.slice1.8.running_var) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:2016:0
%112 : Float(1:5537792, 128:43264, 208:208, 208:1) = onnx::Relu(%111) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:1117:0
%113 : Float(1:5537792, 128:43264, 208:208, 208:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%112, %basenet.slice1.10.weight, %basenet.slice1.10.bias) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/modules/conv.py:416:0
%114 : Float(1:5537792, 128:43264, 208:208, 208:1) = onnx::BatchNormalization[epsilon=1.0000000000000001e-05, momentum=0.90000000000000002](%113, %basenet.slice1.11.weight, %basenet.slice1.11.bias, %basenet.slice1.11.running_mean, %basenet.slice1.11.running_var) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:2016:0
%115 : Float(1:5537792, 128:43264, 208:208, 208:1) = onnx::Relu(%114) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:1117:0
%116 : Float(1:1384448, 128:10816, 104:104, 104:1) = onnx::MaxPoolkernel_shape=[2, 2], pads=[0, 0, 0, 0], strides=[2, 2] # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:576:0
%117 : Float(1:2768896, 256:10816, 104:104, 104:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%116, %basenet.slice2.14.weight, %basenet.slice2.14.bias) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/modules/conv.py:416:0
%118 : Float(1:2768896, 256:10816, 104:104, 104:1) = onnx::BatchNormalization[epsilon=1.0000000000000001e-05, momentum=0.90000000000000002](%117, %basenet.slice2.15.weight, %basenet.slice2.15.bias, %basenet.slice2.15.running_mean, %basenet.slice2.15.running_var) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:2016:0
%119 : Float(1:2768896, 256:10816, 104:104, 104:1) = onnx::Relu(%118) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:1117:0
%120 : Float(1:2768896, 256:10816, 104:104, 104:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%119, %basenet.slice2.17.weight, %basenet.slice2.17.bias) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/modules/conv.py:416:0
%121 : Float(1:2768896, 256:10816, 104:104, 104:1) = onnx::BatchNormalization[epsilon=1.0000000000000001e-05, momentum=0.90000000000000002](%120, %basenet.slice2.18.weight, %basenet.slice2.18.bias, %basenet.slice2.18.running_mean, %basenet.slice2.18.running_var) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:2016:0
%122 : Float(1:2768896, 256:10816, 104:104, 104:1) = onnx::Relu(%121) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:1117:0
%123 : Float(1:2768896, 256:10816, 104:104, 104:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%122, %basenet.slice3.20.weight, %basenet.slice3.20.bias) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/modules/conv.py:416:0
%124 : Float(1:2768896, 256:10816, 104:104, 104:1) = onnx::BatchNormalization[epsilon=1.0000000000000001e-05, momentum=0.90000000000000002](%123, %basenet.slice3.21.weight, %basenet.slice3.21.bias, %basenet.slice3.21.running_mean, %basenet.slice3.21.running_var) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:2016:0
%125 : Float(1:2768896, 256:10816, 104:104, 104:1) = onnx::Relu(%124) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:1117:0
%126 : Float(1:692224, 256:2704, 52:52, 52:1) = onnx::MaxPoolkernel_shape=[2, 2], pads=[0, 0, 0, 0], strides=[2, 2] # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:576:0
%127 : Float(1:1384448, 512:2704, 52:52, 52:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%126, %basenet.slice3.24.weight, %basenet.slice3.24.bias) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/modules/conv.py:416:0
%128 : Float(1:1384448, 512:2704, 52:52, 52:1) = onnx::BatchNormalization[epsilon=1.0000000000000001e-05, momentum=0.90000000000000002](%127, %basenet.slice3.25.weight, %basenet.slice3.25.bias, %basenet.slice3.25.running_mean, %basenet.slice3.25.running_var) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:2016:0
%129 : Float(1:1384448, 512:2704, 52:52, 52:1) = onnx::Relu(%128) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:1117:0
%130 : Float(1:1384448, 512:2704, 52:52, 52:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%129, %basenet.slice3.27.weight, %basenet.slice3.27.bias) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/modules/conv.py:416:0
%131 : Float(1:1384448, 512:2704, 52:52, 52:1) = onnx::BatchNormalization[epsilon=1.0000000000000001e-05, momentum=0.90000000000000002](%130, %basenet.slice3.28.weight, %basenet.slice3.28.bias, %basenet.slice3.28.running_mean, %basenet.slice3.28.running_var) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:2016:0
%132 : Float(1:1384448, 512:2704, 52:52, 52:1) = onnx::Relu(%131) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:1117:0
%133 : Float(1:1384448, 512:2704, 52:52, 52:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%132, %basenet.slice4.30.weight, %basenet.slice4.30.bias) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/modules/conv.py:416:0
%134 : Float(1:1384448, 512:2704, 52:52, 52:1) = onnx::BatchNormalization[epsilon=1.0000000000000001e-05, momentum=0.90000000000000002](%133, %basenet.slice4.31.weight, %basenet.slice4.31.bias, %basenet.slice4.31.running_mean, %basenet.slice4.31.running_var) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:2016:0
%135 : Float(1:1384448, 512:2704, 52:52, 52:1) = onnx::Relu(%134) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:1117:0
%136 : Float(1:346112, 512:676, 26:26, 26:1) = onnx::MaxPoolkernel_shape=[2, 2], pads=[0, 0, 0, 0], strides=[2, 2] # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:576:0
%137 : Float(1:346112, 512:676, 26:26, 26:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%136, %basenet.slice4.34.weight, %basenet.slice4.34.bias) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/modules/conv.py:416:0
%138 : Float(1:346112, 512:676, 26:26, 26:1) = onnx::BatchNormalization[epsilon=1.0000000000000001e-05, momentum=0.90000000000000002](%137, %basenet.slice4.35.weight, %basenet.slice4.35.bias, %basenet.slice4.35.running_mean, %basenet.slice4.35.running_var) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:2016:0
%139 : Float(1:346112, 512:676, 26:26, 26:1) = onnx::Relu(%138) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:1117:0
%140 : Float(1:346112, 512:676, 26:26, 26:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%139, %basenet.slice4.37.weight, %basenet.slice4.37.bias) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/modules/conv.py:416:0
%141 : Float(1:346112, 512:676, 26:26, 26:1) = onnx::BatchNormalization[epsilon=1.0000000000000001e-05, momentum=0.90000000000000002](%140, %basenet.slice4.38.weight, %basenet.slice4.38.bias, %basenet.slice4.38.running_mean, %basenet.slice4.38.running_var) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:2016:0
%142 : Float(1:346112, 512:676, 26:26, 26:1) = onnx::MaxPoolkernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1] # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:576:0
%143 : Float(1:692224, 1024:676, 26:26, 26:1) = onnx::Conv[dilations=[6, 6], group=1, kernel_shape=[3, 3], pads=[6, 6, 6, 6], strides=[1, 1]](%142, %basenet.slice5.1.weight, %basenet.slice5.1.bias) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/modules/conv.py:416:0
%144 : Float(1:692224, 1024:676, 26:26, 26:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%143, %basenet.slice5.2.weight, %basenet.slice5.2.bias) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/modules/conv.py:416:0
%145 : Float(1:1038336, 1536:676, 26:26, 26:1) = onnx::Concat[axis=1](%144, %141) # /home/ec2-user/SageMaker/CRAFT-pytorch/craft.py:63:0
%146 : Float(1:346112, 512:676, 26:26, 26:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%145, %upconv1.conv.0.weight, %upconv1.conv.0.bias) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/modules/conv.py:416:0
%147 : Float(1:346112, 512:676, 26:26, 26:1) = onnx::BatchNormalization[epsilon=1.0000000000000001e-05, momentum=0.90000000000000002](%146, %upconv1.conv.1.weight, %upconv1.conv.1.bias, %upconv1.conv.1.running_mean, %upconv1.conv.1.running_var) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:2016:0
%148 : Float(1:346112, 512:676, 26:26, 26:1) = onnx::Relu(%147) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:1117:0
%149 : Float(1:173056, 256:676, 26:26, 26:1) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%148, %upconv1.conv.3.weight, %upconv1.conv.3.bias) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/modules/conv.py:416:0
%150 : Float(1:173056, 256:676, 26:26, 26:1) = onnx::BatchNormalization[epsilon=1.0000000000000001e-05, momentum=0.90000000000000002](%149, %upconv1.conv.4.weight, %upconv1.conv.4.bias, %upconv1.conv.4.running_mean, %upconv1.conv.4.running_var) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:2016:0
%151 : Float(1:173056, 256:676, 26:26, 26:1) = onnx::Relu(%150) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:1117:0
%152 : Tensor = onnx::Shape(%132)
%153 : Tensor = onnx::Constantvalue={2}
%154 : Long() = onnx::Gather[axis=0](%152, %153) # /home/ec2-user/SageMaker/CRAFT-pytorch/craft.py:66:0
%155 : Tensor = onnx::Shape(%132)
%156 : Tensor = onnx::Constantvalue={3}
%157 : Long() = onnx::Gather[axis=0](%155, %156) # /home/ec2-user/SageMaker/CRAFT-pytorch/craft.py:66:0
%158 : Tensor = onnx::Unsqueezeaxes=[0]
%159 : Tensor = onnx::Unsqueezeaxes=[0]
%160 : Tensor = onnx::Concat[axis=0](%158, %159)
%161 : Tensor = onnx::Constantvalue= 1 1 [ CPUFloatType{2} ]
%162 : Tensor = onnx::Castto=1
%163 : Tensor = onnx::Shape(%151)
%164 : Tensor = onnx::Sliceaxes=[0], ends=[9223372036854775807], starts=[2]
%165 : Tensor = onnx::Castto=1
%166 : Tensor = onnx::Div(%162, %165)
%167 : Tensor = onnx::Concat[axis=0](%161, %166)
%168 : Float(1:692224, 256:2704, 52:52, 52:1) = onnx::Upsample[mode="linear"](%151, %167) # /home/ec2-user/anaconda3/envs/pytorch_latest_p36/lib/python3.6/site-packages/torch/nn/functional.py:3163:0
return (%168)

I'm getting the following error on AWS

ClientError: InputConfiguration: TVM cannot convert ONNX model. Please make sure the framework you selected is correct. <class 'tvm.tir.expr.Any'> has no attribute value

When I tried it on Oct 9, 2020, there was no problem even with Upsample, but as of Nov 12, 2020, the error occurs.
The same thing happens with Resize

Recommend Projects

React

A declarative, efficient, and flexible JavaScript library for building user interfaces.
Vue.js

🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
Typescript

TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
TensorFlow

An Open Source Machine Learning Framework for Everyone
Django

The Web framework for perfectionists with deadlines.
Laravel

A PHP framework for web artisans
D3

Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

javascript

JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
web

Some thing interesting about web. New door for the world.
server

A server is a program made to process requests and deliver data to clients.
Machine learning

Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Visualization

Some thing interesting about visualization, use data art
Game

Some thing interesting about game, make everyone happy.

Recommend Org

Facebook

We are working to build community through open source technology. NB: members must have two-factor auth.
Microsoft

Open source projects and samples from Microsoft.
Google

Google ❤️ Open Source for everyone.
Alibaba

Alibaba Open Source for everyone
D3

Data-Driven Documents codes.
Tencent

China tencent open source team.