hcho3 commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-619255802
@leandron @areusch @tqchen I've put up RC1 for the upcoming XGBoost 1.1.0
release. Feel free to try it:
```
python3 -m pip install xgboost==1.1.0rc1
```
The unit t
hcho3 commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618732386
With https://github.com/dmlc/xgboost/pull/5590, I can now run
`tests/python/unittest/test_autotvm_xgboost_model.py::test_fit` without
crashing.
hcho3 commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618685136
Got it. How about compiling the wheel using latest Ubuntu (not CentOS) and
put it in a S3 bucket? The TVM CI can pull from this bucket instead of PyPI.
Current build enviro
hcho3 commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618680692
@trivialfis Let me file a pull request. We'll include the fix as part of the
upcoming 1.1.0 release.
This is an
hcho3 commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618678730
@trivialfis I think we can hide all C++ symbols when building Python wheels.
WDYT?
This is an automated message
hcho3 commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618677643
@trivialfis Can you elaborate what your patch does? Does it hide certain
symbols?
This is an automated message f
hcho3 commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618673399
@trivialfis I applied your patch and changed CMake flags. And now the unit
test does not crash any more.
You should try it too. Get the wheel at
https://xgboost-wheels.s3-
hcho3 commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618656005
Finally, I reproduced it. Yes!
This is an automated message from the Apache Git Service.
To respond to the messa
hcho3 commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618649425
@areusch Can you try out the latest TVM master on your end? I'm still having
trouble reproducing.
This is an aut
hcho3 commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618644587
@trivialfis And you ran `git submodule update --init --recursive`?
This is an automated message from the Apache G
hcho3 commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618641439
@trivialfis Did you update TVM to latest?
This is an automated message from the Apache Git Service.
To respond to
hcho3 commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618560244
@trivialfis Here it is:
https://drive.google.com/file/d/13WZRRaUPKil4rwH2avgUix_xO5LpIYs5/view?usp=sharing
This
hcho3 commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618188606
@trivialfis I rebuilt the 1.0.2 wheel with debug symbol enabled:
https://drive.google.com/file/d/1cELaBb_rnmb9y8irSwEBQ5AIvaDQgDcF/view?usp=sharing.
Feel free try it.
@are
hcho3 commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-618100776
I am still unable to reproduce it on my Bionic machine. @areusch Can you
share the content of your `config.cmake`?
@trivialfis I'll try to build a wheel using CentOS Docker
hcho3 commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-617492866
Let me try again with `TVM_FFI=ctypes` environment variable set. What does
this do?
This is an automated message
hcho3 commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-616770561
@leandron I checked out commit 8502691b5b7ca152da9eb626529070db53d479c8 so
that XGBoost 1.0.x is used.
This is a
hcho3 commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-616325899
I also tried building `ci-cpu` Docker image from scratch and running the
unit test myself. The test runs without crashing.
--
hcho3 commented on issue #4953:
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-616318728
I compiled TVM from source and tried running the test
`tests/python/unittest/test_autotvm_xgboost_model.py::test_fit` and I cannot
reproduce the issue. Do I need a specific Docke
hcho3 commented on issue #4953: [CI][Docker] xgboost 1.0.1 causes segfault on
test_autotvm_xgboost_model.py
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-610119249
@leandron Thanks for pointing out which part of TVM test is failing. Not
sure if running in debug mode
hcho3 commented on issue #4953: [CI][Docker] xgboost 1.0.1 causes segfault on
test_autotvm_xgboost_model.py
URL: https://github.com/apache/incubator-tvm/issues/4953#issuecomment-592087549
It would be nice if we can get a reproducible example. We are currently
working on the patch release 1
20 matches
Mail list logo