Mujoco-py: Unable to install mujoco-py
Hi,
I have installed mujoco and placed the license file as per the documentation. LD_LIBRARY_PATH is set to $HOME/.mujoco/mjpro150/bin and I am able to run demo such as ./simulate ../model/humanoid.xml.
Now I am trying to install mujoco-py. I am using conda package management system.
Environment details
- Python 3.5.4 |Anaconda, Inc.| (default, Nov 20 2017, 18:44:38)
- conda 4.4.9
- Ubuntu 14.04.5 LTS
- Linux lab 3.13.0-125-generic #174-Ubuntu SMP Mon Jul 10 18:51:24 UTC 2017 x86_64 x86_64 x86_64 GNU/Linux
- Cuda compilation tools, release 7.5, V7.5.17
- 01:00.0 VGA compatible controller [0300]: NVIDIA Corporation GM206 [GeForce GTX 960] [10de:1401] (rev a1) (prog-if 00 [VGA controller])
Installation Steps
Attempt 1
- Activate Python 3.5 Environment i.e.,
source activate python35env cd ~/.mujocogit clone https://github.com/openai/mujoco-py.gitpip install -r requirements.txtpip install -r requirements.dev.txtpython setup.py install
_No error were reported till here_
Use mujoco-py
$ python3
import mujoco_py
I tried to import mujoco_py as shown above. However, the computer starts compiling it and returned after a while. The compilation was failed and following error was reported at terminal-
File "/home/ravi/.pyenv/versions/anaconda3-5.0.1/envs/python35env/lib/python3.5/site-packages/mujoco_py-1.50.1.36-py3.5.egg/mujoco_py/builder.py", line 127, in caching_compile
self._compile(obj, src, ext, cc_args, extra_postargs, pp_opts)
File "/home/ravi/.pyenv/versions/anaconda3-5.0.1/envs/python35env/lib/python3.5/distutils/unixccompiler.py", line 120, in _compile
raise CompileError(msg)
distutils.errors.CompileError: command 'gcc' failed with exit status 1
Attempt 2
I tried pip3 install -U 'mujoco-py<1.50.2,>=1.50.1' but didn't succeed. See the errors below-
[11/11] Cythonizing /tmp/easy_install-9xftm166/Cython-0.27.3/Cython/Tempita/_tempita.py
Unable to find pgen, not compiling formal grammar.
warning: no files found matching '2to3-fixers.txt'
warning: no files found matching 'Doc/*'
warning: no files found matching '*.pyx' under directory 'Cython/Debugger/Tests'
distutils.errors.DistutilsError: Setup script exited with error: Command "x86_64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector --param=ssp-buffer-size=4 -Wformat -Werror=format-security -D_FORTIFY_SOURCE=2 -fPIC -Ibuild/src.linux-x86_64-3.4/numpy/core/src/npymath -Inumpy/core/include -Ibuild/src.linux-x86_64-3.4/numpy/core/include/numpy -I/usr/include/python3.4m -Ibuild/src.linux-x86_64-3.4/numpy/core/src/private -Ibuild/src.linux-x86_64-3.4/numpy/core/src/npymath -Ibuild/src.linux-x86_64-3.4/numpy/core/src/private -Ibuild/src.linux-x86_64-3.4/numpy/core/src/npymath -Ibuild/src.linux-x86_64-3.4/numpy/core/src/private -Ibuild/src.linux-x86_64-3.4/numpy/core/src/npymath -c numpy/core/src/npymath/npy_math.c -o build/temp.linux-x86_64-3.4/numpy/core/src/npymath/npy_math.o -MMD -MF build/temp.linux-x86_64-3.4/numpy/core/src/npymath/npy_math.o.d" failed with exit status 1
In file included from numpy/core/src/npymath/npy_math.c:9:0:
numpy/core/src/npymath/npy_math_internal.h.src:56:30: fatal error: npy_math_private.h: No such file or directory
#include "npy_math_private.h"
^
compilation terminated.
In file included from numpy/core/src/npymath/npy_math.c:9:0:
numpy/core/src/npymath/npy_math_internal.h.src:56:30: fatal error: npy_math_private.h: No such file or directory
#include "npy_math_private.h"
^
compilation terminated.
----------------------------------------
Cleaning up...
Command python setup.py egg_info failed with error code 1 in /tmp/pip_build_ravi/mujoco-py
Storing debug log for failure in /home/ravi/.pip/pip.log
Additional Information
Since above error is related to numpy, so I verified it. I am using numpy 1.13.3 as shown below-
(python35env) ravi@lab:~$ conda list | grep numpy
numpy 1.13.3 py35h3dfced4_2
numpydoc 0.7.0 <pip>
Attempt 3
I also tried pip install mujoco-py but no success. See the errors below-
File "/tmp/pip-build-8wxj6ah8/mujoco-py/setup.py", line 19, in run
spec.loader.exec_module(builder)
File "<frozen importlib._bootstrap_external>", line 697, in exec_module
File "<frozen importlib._bootstrap>", line 222, in _call_with_frames_removed
File "/tmp/pip-build-8wxj6ah8/mujoco-py/mujoco_py/builder.py", line 22, in <module>
from mujoco_py.utils import discover_mujoco
File "/tmp/pip-build-8wxj6ah8/mujoco-py/mujoco_py/__init__.py", line 1, in <module>
from mujoco_py.builder import cymj, ignore_mujoco_warnings, functions, MujocoException
File "/tmp/pip-build-8wxj6ah8/mujoco-py/mujoco_py/builder.py", line 502, in <module>
cymj = load_cython_ext(mjpro_path)
File "/tmp/pip-build-8wxj6ah8/mujoco-py/mujoco_py/builder.py", line 82, in load_cython_ext
return load_dynamic_ext('cymj', cext_so_path)
File "/tmp/pip-build-8wxj6ah8/mujoco-py/mujoco_py/builder.py", line 88, in load_dynamic_ext
return loader.load_module()
ImportError: /usr/local/nvidia/lib64/libEGL.so.1: cannot open shared object file: No such file or directory
----------------------------------------
Command "/home/ravi/.pyenv/versions/anaconda3-5.0.1/envs/python35env/bin/python -u -c "import setuptools, tokenize;__file__='/tmp/pip-build-8wxj6ah8/mujoco-py/setup.py';f=getattr(tokenize, 'open', open)(__file__);code=f.read().replace('\r\n', '\n');f.close();exec(compile(code, __file__, 'exec'))" install --record /tmp/pip-bdwvwc34-record/install-record.txt --single-version-externally-managed --compile" failed with error code 1 in /tmp/pip-build-8wxj6ah8/mujoco-py/
Additional Information
This time, it seems that the error is related to nvidia. Unfortunately, there is no such file /usr/local/nvidia/lib64/libEGL.so.1
ravi@lab:~$ ls -la /usr/local/nvidia/lib64/libEGL.so.1
ls: cannot access /usr/local/nvidia/lib64/libEGL.so.1: No such file or directory
ravi@lab:~$ ls -la /usr/local/nvidia
ls: cannot access /usr/local/nvidia: No such file or directory
I am confused!!! What is the problem exactly? Any workaround, please?
ravijo
All 35 comments
I think python 3.5.4 is not supported and you should use 3.5.2
eliasdc
on 12 Feb 2018
@eliasdc: Thank you very much. I created new conda environment with python 3.5.2 and tried installing it again but didn't succeed. Below is the error-
raise DistutilsError("Setup script exited with %s" % (v.args[0],))
distutils.errors.DistutilsError: Setup script exited with error: Command "x86_64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector --param=ssp-buffer-size=4 -Wformat -Werror=format-security -D_FORTIFY_SOURCE=2 -fPIC -Ibuild/src.linux-x86_64-3.4/numpy/core/src/npymath -Inumpy/core/include -Ibuild/src.linux-x86_64-3.4/numpy/core/include/numpy -I/usr/include/python3.4m -Ibuild/src.linux-x86_64-3.4/numpy/core/src/private -Ibuild/src.linux-x86_64-3.4/numpy/core/src/npymath -Ibuild/src.linux-x86_64-3.4/numpy/core/src/private -Ibuild/src.linux-x86_64-3.4/numpy/core/src/npymath -Ibuild/src.linux-x86_64-3.4/numpy/core/src/private -Ibuild/src.linux-x86_64-3.4/numpy/core/src/npymath -c build/src.linux-x86_64-3.4/numpy/core/src/npymath/npy_math_complex.c -o build/temp.linux-x86_64-3.4/build/src.linux-x86_64-3.4/numpy/core/src/npymath/npy_math_complex.o -MMD -MF build/temp.linux-x86_64-3.4/build/src.linux-x86_64-3.4/numpy/core/src/npymath/npy_math_complex.o.d" failed with exit status 1
numpy/core/src/npymath/npy_math_complex.c.src:34:29: fatal error: npy_math_common.h: No such file or directory
#include "npy_math_common.h"
^
compilation terminated.
numpy/core/src/npymath/npy_math_complex.c.src:34:29: fatal error: npy_math_common.h: No such file or directory
#include "npy_math_common.h"
^
compilation terminated.
----------------------------------------
Cleaning up...
Command python setup.py egg_info failed with error code 1 in /tmp/pip_build_ravi/mujoco-py
Storing debug log for failure in /home/ravi/.pip/pip.log
Following packages are installed in this environment-
(mujocoenv) ravi@lab:~/.mujoco/mujoco-py$ conda list
# packages in environment at /home/ravi/.pyenv/versions/anaconda3-5.0.1/envs/mujocoenv:
#
# Name Version Build Channel
alabaster 0.7.10 <pip>
attrs 17.4.0 <pip>
Babel 2.5.3 <pip>
ca-certificates 2017.08.26 h1d4fec5_0
certifi 2018.1.18 py35_0
cffi 1.11.4 <pip>
chardet 3.0.4 <pip>
Cython 0.27.3 <pip>
decorator 4.2.1 <pip>
docutils 0.14 <pip>
glfw 1.5.1 <pip>
idna 2.6 <pip>
ImageHash 4.0 <pip>
imageio 2.2.0 <pip>
imagesize 0.7.1 <pip>
intel-openmp 2018.0.0 hc7b2577_8
ipdb 0.10.3 <pip>
ipython 6.2.1 <pip>
ipython-genutils 0.2.0 <pip>
jedi 0.11.1 <pip>
Jinja2 2.10 <pip>
libgcc-ng 7.2.0 h7cc24e2_2
libgfortran-ng 7.2.0 h9f7466a_2
libstdcxx-ng 7.2.0 h7a57d05_2
MarkupSafe 1.0 <pip>
mkl 2018.0.1 h19d6760_4
numpy 1.14.0 py35h3dfced4_1
numpydoc 0.7.0 <pip>
openssl 1.0.2n hb7f436b_0
packaging 16.8 <pip>
parso 0.1.1 <pip>
pexpect 4.4.0 <pip>
pickleshare 0.7.4 <pip>
Pillow 5.0.0 <pip>
pip 9.0.1 py35h7e7da9d_4
pluggy 0.6.0 <pip>
prompt-toolkit 1.0.15 <pip>
ptyprocess 0.5.2 <pip>
py 1.5.2 <pip>
pycparser 2.18 <pip>
Pygments 2.2.0 <pip>
pyparsing 2.2.0 <pip>
pytest 3.4.0 <pip>
pytest-instafail 0.3.0 <pip>
python 3.5.2 0
pytz 2018.3 <pip>
PyWavelets 0.5.2 <pip>
readline 6.2 2
requests 2.18.4 <pip>
scipy 1.0.0 py35hcbbe4a2_0
setuptools 38.4.0 py35_0
simplegeneric 0.8.1 <pip>
six 1.11.0 <pip>
snowballstemmer 1.2.1 <pip>
Sphinx 1.7.0 <pip>
sphinx-rtd-theme 0.2.4 <pip>
sphinxcontrib-websupport 1.0.1 <pip>
sqlite 3.13.0 0
tk 8.5.18 0
traitlets 4.3.2 <pip>
urllib3 1.22 <pip>
wcwidth 0.1.7 <pip>
wheel 0.30.0 py35hd3883cf_1
xz 5.2.3 h55aa19d_2
zlib 1.2.11 ha838bed_2
try set your to $LD_LIBRARY_PATH:$HOME/.mujoco/mjpro150/bin
export LD_LIBRARY_PATH=“$LD_LIBRARY_PATH:$HOME/.mujoco/mjpro150/bin“
If you do
export LD_LIBRARY_PATH=“$HOME/.mujoco/mjpro150/bin@
It will overwrite the previous ld library path, that’s why you can not compile.
CrazyHeex
on 12 Feb 2018
@CrazyHeex: I set LD_LIBRARY_PATH inside .bashrc file as following-
# mujoco configurations
LD_LIBRARY_PATH=${HOME}/.mujoco/mjpro150/bin:${LD_LIBRARY_PATH}
export LD_LIBRARY_PATH
Later, I needed to run (mujocoenv) ravi@lab:~/.mujoco/mujoco-py$ sudo pip3 install -e . with sudo, since it failed once showing permission issues. It took some time to compile and returned with following message on terminal-
To add a missing option, make sure you have the required
library and headers.
See https://pillow.readthedocs.io/en/latest/installation.html#building-from-source
To check the build, run the selftest.py script.
Running setup.py install for pycparser
warning: no previously-included files matching 'yacctab.*' found under directory 'tests'
warning: no previously-included files matching 'lextab.*' found under directory 'tests'
warning: no previously-included files matching 'yacctab.*' found under directory 'examples'
warning: no previously-included files matching 'lextab.*' found under directory 'examples'
Build the lexing/parsing tables
Successfully installed glfw numpy Cython imageio cffi mujoco-py pillow pycparser
Cleaning up...
However, import mujoco_py is still failing with the following message-
File "/home/ravi/.pyenv/versions/anaconda3-5.0.1/envs/mujocoenv/lib/python3.5/distutils/command/build_ext.py", line 532, in build_extension
depends=ext.depends)
File "/home/ravi/.mujoco/mujoco-py/mujoco_py/builder.py", line 127, in caching_compile
self._compile(obj, src, ext, cc_args, extra_postargs, pp_opts)
File "/home/ravi/.pyenv/versions/anaconda3-5.0.1/envs/mujocoenv/lib/python3.5/distutils/unixccompiler.py", line 120, in _compile
raise CompileError(msg)
distutils.errors.CompileError: command 'gcc' failed with exit status 1
I get to my laptop and did following command with no problem:
conda create -n test python=3.5.2
git clone https://github.com/openai/mujoco-py
pip install -e ./mujoco-py
python
import mujoco_py
conda list
packages in environment at /home/zyang/anaconda3/envs/test:
ca-certificates 2017.08.26 h1d4fec5_0
certifi 2018.1.18 py35_0
cffi 1.11.4
Cython 0.27.3
glfw 1.5.1
imageio 2.2.0
libgcc-ng 7.2.0 h7cc24e2_2
numpy 1.14.0
openssl 1.0.2n hb7f436b_0
Pillow 5.0.0
pip 9.0.1 py35h7e7da9d_4
pycparser 2.18
python 3.5.2 0
readline 6.2 2
setuptools 38.4.0 py35_0
sqlite 3.13.0 0
tk 8.5.18 0
wheel 0.30.0 py35hd3883cf_1
xz 5.2.3 h55aa19d_2
zlib 1.2.11 ha838bed_2
btw,
try:
conda install -c clinicalgraphics osmesa
conda update
idk why but i found this in the history of my cluster when I try install mujoco-py there, i forgot why i did this but may help
CrazyHeex
on 12 Feb 2018
@CrazyHeex: Still no success. Just copied your steps as following-
conda create -n test python=3.5.2source activate testcd ~/.mujocogit clone https://github.com/openai/mujoco-pyexport LD_LIBRARY_PATH=“$LD_LIBRARY_PATH:$HOME/.mujoco/mjpro150/bin“// although it is in.bashrcfile toopip install -e ./mujoco-py
See below the output in terminal-
(test) ravi@lab:~/.mujoco$ pip install -e ./mujoco-py
Obtaining file:///home/ravi/.mujoco/mujoco-py
Collecting glfw>=1.4.0 (from mujoco-py==1.50.1.36)
Collecting numpy>=1.11 (from mujoco-py==1.50.1.36)
Downloading numpy-1.14.0-cp35-cp35m-manylinux1_x86_64.whl (17.1MB)
100% |████████████████████████████████| 17.1MB 111kB/s
Collecting Cython>=0.27.2 (from mujoco-py==1.50.1.36)
Using cached Cython-0.27.3-cp35-cp35m-manylinux1_x86_64.whl
Collecting imageio>=2.1.2 (from mujoco-py==1.50.1.36)
Collecting cffi>=1.10 (from mujoco-py==1.50.1.36)
Using cached cffi-1.11.4-cp35-cp35m-manylinux1_x86_64.whl
Collecting pillow (from imageio>=2.1.2->mujoco-py==1.50.1.36)
Using cached Pillow-5.0.0-cp35-cp35m-manylinux1_x86_64.whl
Collecting pycparser (from cffi>=1.10->mujoco-py==1.50.1.36)
Installing collected packages: glfw, numpy, Cython, pillow, imageio, pycparser, cffi, mujoco-py
Running setup.py develop for mujoco-py
Successfully installed Cython-0.27.3 cffi-1.11.4 glfw-1.5.1 imageio-2.2.0 mujoco-py numpy-1.14.0 pillow-5.0.0 pycparser-2.18
Now see the errors while trying import mujoco_py below-
(test) ravi@lab:~/.mujoco$ python
Python 3.5.2 |Continuum Analytics, Inc.| (default, Jul 2 2016, 17:53:06)
[GCC 4.4.7 20120313 (Red Hat 4.4.7-1)] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import mujoco_py
Compiling /home/ravi/.mujoco/mujoco-py/mujoco_py/cymj.pyx because it changed.
[1/1] Cythonizing /home/ravi/.mujoco/mujoco-py/mujoco_py/cymj.pyx
running build_ext
building 'mujoco_py.cymj' extension
creating /home/ravi/.mujoco/mujoco-py/mujoco_py/generated/_pyxbld_LinuxExtensionBuilder
creating /home/ravi/.mujoco/mujoco-py/mujoco_py/generated/_pyxbld_LinuxExtensionBuilder/temp.linux-x86_64-3.5
creating /home/ravi/.mujoco/mujoco-py/mujoco_py/generated/_pyxbld_LinuxExtensionBuilder/temp.linux-x86_64-3.5/home
creating /home/ravi/.mujoco/mujoco-py/mujoco_py/generated/_pyxbld_LinuxExtensionBuilder/temp.linux-x86_64-3.5/home/ravi
creating /home/ravi/.mujoco/mujoco-py/mujoco_py/generated/_pyxbld_LinuxExtensionBuilder/temp.linux-x86_64-3.5/home/ravi/.mujoco
creating /home/ravi/.mujoco/mujoco-py/mujoco_py/generated/_pyxbld_LinuxExtensionBuilder/temp.linux-x86_64-3.5/home/ravi/.mujoco/mujoco-py
creating /home/ravi/.mujoco/mujoco-py/mujoco_py/generated/_pyxbld_LinuxExtensionBuilder/temp.linux-x86_64-3.5/home/ravi/.mujoco/mujoco-py/mujoco_py
creating /home/ravi/.mujoco/mujoco-py/mujoco_py/generated/_pyxbld_LinuxExtensionBuilder/temp.linux-x86_64-3.5/home/ravi/.mujoco/mujoco-py/mujoco_py/gl
gcc -pthread -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -fPIC -Imujoco-py/mujoco_py -I/home/ravi/.mujoco/mujoco-py/mujoco_py -I/home/ravi/.mujoco/mjpro150/include -I/home/ravi/.pyenv/versions/anaconda3-5.0.1/envs/test/lib/python3.5/site-packages/numpy/core/include -I/home/ravi/.mujoco/mujoco-py/mujoco_py/vendor/egl -I/home/ravi/.pyenv/versions/anaconda3-5.0.1/envs/test/include/python3.5m -c /home/ravi/.mujoco/mujoco-py/mujoco_py/cymj.c -o /home/ravi/.mujoco/mujoco-py/mujoco_py/generated/_pyxbld_LinuxExtensionBuilder/temp.linux-x86_64-3.5/home/ravi/.mujoco/mujoco-py/mujoco_py/cymj.o -fopenmp -w
gcc -pthread -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -fPIC -Imujoco-py/mujoco_py -I/home/ravi/.mujoco/mujoco-py/mujoco_py -I/home/ravi/.mujoco/mjpro150/include -I/home/ravi/.pyenv/versions/anaconda3-5.0.1/envs/test/lib/python3.5/site-packages/numpy/core/include -I/home/ravi/.mujoco/mujoco-py/mujoco_py/vendor/egl -I/home/ravi/.pyenv/versions/anaconda3-5.0.1/envs/test/include/python3.5m -c /home/ravi/.mujoco/mujoco-py/mujoco_py/gl/eglshim.c -o /home/ravi/.mujoco/mujoco-py/mujoco_py/generated/_pyxbld_LinuxExtensionBuilder/temp.linux-x86_64-3.5/home/ravi/.mujoco/mujoco-py/mujoco_py/gl/eglshim.o -fopenmp -w
/home/ravi/.mujoco/mujoco-py/mujoco_py/gl/eglshim.c: In function ‘closeOpenGL’:
/home/ravi/.mujoco/mujoco-py/mujoco_py/gl/eglshim.c:159:5: error: ‘for’ loop initial declarations are only allowed in C99 mode
for (int device_id=0; device_id<MAX_DEVICES; device_id++) {
^
/home/ravi/.mujoco/mujoco-py/mujoco_py/gl/eglshim.c:159:5: note: use option -std=c99 or -std=gnu99 to compile your code
Traceback (most recent call last):
File "/home/ravi/.pyenv/versions/anaconda3-5.0.1/envs/test/lib/python3.5/distutils/unixccompiler.py", line 118, in _compile
extra_postargs)
File "/home/ravi/.pyenv/versions/anaconda3-5.0.1/envs/test/lib/python3.5/distutils/ccompiler.py", line 909, in spawn
spawn(cmd, dry_run=self.dry_run)
File "/home/ravi/.pyenv/versions/anaconda3-5.0.1/envs/test/lib/python3.5/distutils/spawn.py", line 36, in spawn
_spawn_posix(cmd, search_path, dry_run=dry_run)
File "/home/ravi/.pyenv/versions/anaconda3-5.0.1/envs/test/lib/python3.5/distutils/spawn.py", line 159, in _spawn_posix
% (cmd, exit_status))
distutils.errors.DistutilsExecError: command 'gcc' failed with exit status 1
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/home/ravi/.mujoco/mujoco-py/mujoco_py/__init__.py", line 1, in <module>
from mujoco_py.builder import cymj, ignore_mujoco_warnings, functions, MujocoException
File "/home/ravi/.mujoco/mujoco-py/mujoco_py/builder.py", line 502, in <module>
cymj = load_cython_ext(mjpro_path)
File "/home/ravi/.mujoco/mujoco-py/mujoco_py/builder.py", line 81, in load_cython_ext
cext_so_path = builder.build()
File "/home/ravi/.mujoco/mujoco-py/mujoco_py/builder.py", line 238, in build
built_so_file_path = self._build_impl()
File "/home/ravi/.mujoco/mujoco-py/mujoco_py/builder.py", line 310, in _build_impl
so_file_path = super()._build_impl()
File "/home/ravi/.mujoco/mujoco-py/mujoco_py/builder.py", line 261, in _build_impl
dist.run_commands()
File "/home/ravi/.pyenv/versions/anaconda3-5.0.1/envs/test/lib/python3.5/distutils/dist.py", line 955, in run_commands
self.run_command(cmd)
File "/home/ravi/.pyenv/versions/anaconda3-5.0.1/envs/test/lib/python3.5/distutils/dist.py", line 974, in run_command
cmd_obj.run()
File "/home/ravi/.pyenv/versions/anaconda3-5.0.1/envs/test/lib/python3.5/site-packages/Cython/Distutils/old_build_ext.py", line 186, in run
_build_ext.build_ext.run(self)
File "/home/ravi/.pyenv/versions/anaconda3-5.0.1/envs/test/lib/python3.5/distutils/command/build_ext.py", line 338, in run
self.build_extensions()
File "/home/ravi/.mujoco/mujoco-py/mujoco_py/builder.py", line 157, in build_extensions
build_ext.build_extensions(self)
File "/home/ravi/.pyenv/versions/anaconda3-5.0.1/envs/test/lib/python3.5/site-packages/Cython/Distutils/old_build_ext.py", line 194, in build_extensions
self.build_extension(ext)
File "/home/ravi/.pyenv/versions/anaconda3-5.0.1/envs/test/lib/python3.5/distutils/command/build_ext.py", line 532, in build_extension
depends=ext.depends)
File "/home/ravi/.mujoco/mujoco-py/mujoco_py/builder.py", line 127, in caching_compile
self._compile(obj, src, ext, cc_args, extra_postargs, pp_opts)
File "/home/ravi/.pyenv/versions/anaconda3-5.0.1/envs/test/lib/python3.5/distutils/unixccompiler.py", line 120, in _compile
raise CompileError(msg)
distutils.errors.CompileError: command 'gcc' failed with exit status 1
>>>
Here is the list of packages in my conda environment-
(test) ravi@lab:~/.mujoco$ conda list
# packages in environment at /home/ravi/.pyenv/versions/anaconda3-5.0.1/envs/test:
#
# Name Version Build Channel
ca-certificates 2017.08.26 h1d4fec5_0
certifi 2018.1.18 py35_0
cffi 1.11.4 <pip>
Cython 0.27.3 <pip>
glfw 1.5.1 <pip>
imageio 2.2.0 <pip>
libgcc-ng 7.2.0 h7cc24e2_2
mujoco-py 1.50.1.36 <pip>
numpy 1.14.0 <pip>
openssl 1.0.2n hb7f436b_0
Pillow 5.0.0 <pip>
pip 9.0.1 py35h7e7da9d_4
pycparser 2.18 <pip>
python 3.5.2 0
readline 6.2 2
setuptools 38.4.0 py35_0
sqlite 3.13.0 0
tk 8.5.18 0
wheel 0.30.0 py35hd3883cf_1
xz 5.2.3 h55aa19d_2
zlib 1.2.11 ha838bed_2
Additional info
conda install -c clinicalgraphics osmesa failed by showing PackagesNotFoundError. Then I installed conda install -c menpo osmesa. Please see below-
(test) ravi@lab:~/.mujoco$ conda install -c clinicalgraphics osmesa
Solving environment: failed
PackagesNotFoundError: The following packages are not available from current channels:
- osmesa
Current channels:
- https://conda.anaconda.org/clinicalgraphics/linux-64
- https://conda.anaconda.org/clinicalgraphics/noarch
- https://repo.continuum.io/pkgs/main/linux-64
- https://repo.continuum.io/pkgs/main/noarch
- https://repo.continuum.io/pkgs/free/linux-64
- https://repo.continuum.io/pkgs/free/noarch
- https://repo.continuum.io/pkgs/r/linux-64
- https://repo.continuum.io/pkgs/r/noarch
- https://repo.continuum.io/pkgs/pro/linux-64
- https://repo.continuum.io/pkgs/pro/noarch
(test) ravi@lab:~/.mujoco$ conda install -c menpo osmesa
Solving environment: done
Downloading and Extracting Packages
osmesa 12.2.2.dev: ##################################################### | 100%
Preparing transaction: done
Verifying transaction: done
Executing transaction: done
Unfortunately conda update didn't work. Please see below-
(test) ravi@lab:~/.mujoco$ conda update
CondaValueError: no package names supplied
# If you want to update to a newer version of Anaconda, type:
#
# $ conda update --prefix /home/ravi/.pyenv/versions/anaconda3-5.0.1/envs/test anaconda
try to move the int out of the for loop in the code:/home/ravi/.mujoco/mujoco-py/mujoco_py/gl/eglshim.c:159:5
for (int device_id=0; device_id to int device_id = 0; not sure, I never saw this before.
for (;device_id
CrazyHeex
on 12 Feb 2018
int device_id;
for (device_id=0; device_id<MAX_DEVICES; device_id++)
It means that the compiler is not using -std=c++11. Isn't it suspicious? Also the above code resides inside OpenGL function. Is the error related to OpenGL?
The above works fine and I am able to use import mujoco_py as shown below-
>>> import mujoco_py
>>> from os.path import dirname
>>> model = mujoco_py.load_model_from_path(dirname(dirname(mujoco_py.__file__)) +"/xmls/claw.xml")
>>> sim = mujoco_py.MjSim(model)
>>> print(sim.data.qpos)
[0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.]
However, when I treid to use gym, I got following error ERROR: GLEW initalization error: Missing GL version
(test) ravi@lab:~/.mujoco/mjpro150/bin$ python3
Python 3.5.2 |Continuum Analytics, Inc.| (default, Jul 2 2016, 17:53:06)
[GCC 4.4.7 20120313 (Red Hat 4.4.7-1)] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import gym
>>> env = gym.make('Hopper-v2')
WARN: gym.spaces.Box autodetected dtype as <class 'numpy.float32'>. Please provide explicit dtype.
WARN: gym.spaces.Box autodetected dtype as <class 'numpy.float32'>. Please provide explicit dtype.
>>> print(env.observation_space)
Box(11,)
>>> print(env.action_space)
Box(3,)
>>> env.reset()
array([ 1.24700988e+00, -4.12321754e-03, 3.67075481e-03, -2.96215107e-03,
-1.26082879e-03, 5.48675500e-04, -4.46060702e-03, 2.28999331e-03,
-2.64810316e-03, -2.33473989e-03, -4.64960736e-03])
>>> for _ in range(100):
... env.render()
... env.step(env.action_space.sample()) # take a random action
...
Creating window glfw
ERROR: GLEW initalization error: Missing GL version
Press Enter to exit ...Killed
Whats' wrong now? Quick search is showing me that the error can be because of libglewosmesa. Should I remove libglewosmesa and try again?
Please note that I am still able to use ./simulate ../model/humanoid.xml and can see the humanoid simulation but when I close this window, it shows Segmentation fault. See Below-
(test) ravi@lab:~/.mujoco/mjpro150/bin$ ./simulate ../model/humanoid.xmlMuJoCo Pro library version 1.50
Segmentation fault
Did you sort this? I'm having the same problem.
JamesOwers
on 13 Feb 2018
I was having the same problems and decided to use the older mujoco-py in combination with an older openai gym version.
gym==0.9.3
mujoco-py==0.5.7
mujoco 131 from their website
eliasdc
on 15 Feb 2018
@eliasdc: As per your suggestion, I will give it a try. However, I am worried about the license file. Does the same license file work for mujoco v131 too? I have the license file for mujoco v150.
ravijo
on 15 Feb 2018
@ravijo: The same license file can be used.
eliasdc
on 15 Feb 2018
As a workaround suggest by @eliasdc, I found that following combination works fine-
- gym==0.9.3
- mujoco-py==0.5.7
- mujoco 131
Below is the list of packages present in the environment while testing-
# Name Version Build Channel
ca-certificates 2017.08.26 h1d4fec5_0
certifi 2018.1.18 py27_0
chardet 3.0.4 <pip>
future 0.16.0 <pip>
gym 0.9.3 <pip>
idna 2.6 <pip>
libedit 3.1 heed3624_0
libffi 3.2.1 hd88cf55_4
libgcc-ng 7.2.0 h7cc24e2_2
libstdcxx-ng 7.2.0 h7a57d05_2
mujoco-py 0.5.7 <pip>
ncurses 6.0 h9df7e31_2
numpy 1.14.0 <pip>
openssl 1.0.2n hb7f436b_0
pip 9.0.1 py27ha730c48_4
pyglet 1.3.1 <pip>
PyOpenGL 3.1.0 <pip>
python 2.7.14 h1571d57_29
readline 7.0 ha6073c6_4
requests 2.18.4 <pip>
setuptools 38.4.0 py27_0
six 1.11.0 <pip>
sqlite 3.22.0 h1bed415_0
tk 8.6.7 hc745277_3
urllib3 1.22 <pip>
wheel 0.30.0 py27h2bc6bb2_1
zlib 1.2.11 ha838bed_2
How do I install gym=0.9.3
If I do pip install gym==0.9.3 it installs 0.10.3
ahsteven
on 2 Mar 2018
@ahsteven: strange! Please make sure you are using Python 2.7. I made a conda environment with Python 2.7 and then installed gym 0.9.3
ravijo
on 2 Mar 2018
Ok I was able to get this installed with py 2.7 and import mujoco_py.
One question, how do I set the version when installing from the github? I installed using pip install gym==0.9.3 but I would like to install from the github with 'pip install -e .' or something similar.
My goal in installing this is to be able to run the openai gym baselines. Which I am getting an invalid syntax error right now so I am thinking the gym baselines might work with py 35.
File "baselines/common/dataset.py", line 50
def iterbatches(arrays, *, num_batches=None, batch_size=None, shuffle=True, include_final_partial_batch=True):
^
SyntaxError: invalid syntax
I will try this again with python 352.
This forum has been very helpful.
ahsteven
on 2 Mar 2018
@ahsteven: checkout the correct tag or commit and than use pip install -e .
eliasdc
on 5 Mar 2018
So the only workaround is to revert to an older version?. Has any one successfully installed mujoco-py v1.5?
SiddhardhaSaran
on 9 Mar 2018
@SiddhardhaSaran: I think it is possible to install. However, in my case, I have CUDA installed which is creating issues. You can try on your system without CUDA.
ravijo
on 9 Mar 2018
running sudo apt-get install libglew-dev
and installing patchelf. I was able to install mujoco v1.5.
SiddhardhaSaran
on 9 Mar 2018
@SiddhardhaSaran I am having the same / similar problem as @ravijo. I had previously installed libglew-dev but did not have patchelf. After I installed patchelf I tried installing again, and it was successful, but it fails on import.
I found this python wrapper works with my GPU (nvidia) and mjpro150: mujoco
lbhdc
on 10 Mar 2018
I have both libglew-dev and patchelf installed in my systems and still during installation I see the missing GLEW.h error:
/tmp/pip-build-7uw627tn/mujoco-py/mujoco_py/gl/eglshim.c:4:21: fatal error: GL/glew.h: No such file or directory
compilation terminated.
error: command '/usr/bin/gcc-5' failed with exit status 1
----------------------------------------
Failed building wheel for mujoco-py
Running setup.py clean for mujoco-py
Failed to build mujoco-py
Installing collected packages: mujoco-py
Running setup.py install for mujoco-py ... error
Complete output from command /home/user/miniconda3/envs/tf/bin/python -u -c "import setuptools, tokenize;__file__='/tmp/pip-build-7uw627tn/mujoco-py/setup.py';f=getattr(tokenize, 'open', open)(__file__);code=f.read().replace('\r\n', '\n');f.close();exec(compile(code, __file__, 'exec'))" install --record /tmp/pip-n5yr5kcp-record/install-record.txt --single-version-externally-managed --compile:
running install
running build
running build_ext
building 'mujoco_py.cymj' extension
/tmp/pip-build-7uw627tn/mujoco-py/mujoco_py/generated/_pyxbld_LinuxExtensionBuilder/temp.linux-x86_64-3.6/tmp/pip-build-7uw627tn/mujoco-py/mujoco_py/cymj.o up to date, skipping.
/usr/bin/gcc-5 -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -fPIC -Imujoco_py -I/tmp/pip-build-7uw627tn/mujoco-py/mujoco_py -I/home/user/.mujoco/mjpro150/include -I/home/user/miniconda3/envs/tf/lib/python3.6/site-packages/numpy/core/include -I/tmp/pip-build-7uw627tn/mujoco-py/mujoco_py/vendor/egl -I/home/user/miniconda3/envs/tf/include/python3.6m -c /tmp/pip-build-7uw627tn/mujoco-py/mujoco_py/gl/eglshim.c -o /tmp/pip-build-7uw627tn/mujoco-py/mujoco_py/generated/_pyxbld_LinuxExtensionBuilder/temp.linux-x86_64-3.6/tmp/pip-build-7uw627tn/mujoco-py/mujoco_py/gl/eglshim.o -fopenmp -w
/tmp/pip-build-7uw627tn/mujoco-py/mujoco_py/gl/eglshim.c:4:21: fatal error: GL/glew.h: No such file or directory
compilation terminated.
error: command '/usr/bin/gcc-5' failed with exit status 1
----------------------------------------
Command "/home/user/miniconda3/envs/tf/bin/python -u -c "import setuptools, tokenize;__file__='/tmp/pip-build-7uw627tn/mujoco-py/setup.py';f=getattr(tokenize, 'open', open)(__file__);code=f.read().replace('\r\n', '\n');f.close();exec(compile(code, __file__, 'exec'))" install --record /tmp/pip-n5yr5kcp-record/install-record.txt --single-version-externally-managed --compile" failed with error code 1 in /tmp/pip-build-7uw627tn/mujoco-py/
Everything that I've seen so far here is just a hack to make mujoco-py work, we need true support of python 3.6.x
kirk86
on 14 Mar 2018
Everything that I've seen so far here is just a hack to make mujoco-py work, we need true support of python 3.6.x
@kirk86 : Please make sure you are using Python 3.5.2. Support for Python 3.6 is planned
ravijo
on 15 Mar 2018
I'm using python 3.5.2, but still got exactly the same error. (Ubuntu)
YaqiXie
on 20 Mar 2018
@YaqiXie: I would suggest to setup following environment:
- python v2.7.14
- gym v 0.9.3
- mujoco v131
- mujoco-py v0.5.7
Check out this for more info https://github.com/openai/mujoco-py/issues/198#issuecomment-366200529
ravijo
on 21 Mar 2018
I have similar issue with installing mujoco as well.
For me every things is installed and located under the python virtual environment.
The pip install mujoco-py==0.5.7 works very well. But when is the turn for the
pip3 install -U 'mujoco-py<1.50.2,>=1.50.1'
I get the error of:
Exception: Please add path to mujoco library to your .bashrc:
export DYLD_LIBRARY_PATH=$DYLD_LIBRARY_PATH:/Users/user/.mujoco/mjpro150/bin
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/Users/user.mujoco/mjpro150/bin
While I have in my bashrc the path to the virtual py:
export DYLD_LIBRARY_PATH=$DYLD_LIBRARY_PATH:/Users/user/.pyenv/versions/3.6.1/envs/py_env_azi/lib/python3.6/site-packages/gym/envs/mujoco/mjpro150/bin
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/Users/user/.pyenv/versions/3.6.1/envs/py_env_azi/lib/python3.6/site-packages/gym/envs/mujoco/mjpro150/bin
So, I dont know why it can not figure it out the location of binary files although I have the correct path under my bashrc!
azmosavi
on 5 Apr 2018
@azmosavi: It seems that you are on Mac. Unfortunately, I haven't tried MuJoCo on mac. A quick glance at the documentation of mujoco-py says following about MuJoCo installation:
Unzip the downloaded mjpro150 directory into ~/.mujoco/mjpro150
Hence, here in my Ubuntu 14.04 LTS, I place MuJoCo binaries in $HOME. I would suggest you do the same on your system.
ravijo
on 5 Apr 2018
Thanks @ravijo. I solved my problem. Only having --no-cache at the end maid the miracle
git clone https://github.com/openai/mujoco-py
cd mujoco-py
pip install -e . --no-cache
azmosavi
on 5 Apr 2018
This might be a completely wrong answer, but I was having issues installing mujoco-py myself with similar errors to yours and doing:
sudo apt-get install libglew1.5
and
sudo apt-get install libglew-dev
and putting the mjkey.txt in the .mujoco/ folder
and running chmod +x on mjkey.txt
I'm not sure which of these solved my problem, but I was able to import after that. I'm using the new version for python3.6
tensorcoder
on 11 Aug 2018
Worked!!!
In terminal:
sudo apt-get install -y python-numpy cmake zlib1g-dev libjpeg-dev libboost-all-dev gcc libsdl2-dev wget unzip
sudo pip3 install openai gym tensorflow
In forder mujoco_py:
sudo pip3 install -e . --no-cache
In .mujoco/ add key.txt and mujoco unzip
Export path to ~/.bashrc
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/home/(name_user)/.mujoco/mjpro150/bin
damvantai
on 22 Nov 2018
Closing this issue for now because the build system has changed since.
If any new users encounter something similar, please open a new issue! There's an issue template which helps us in getting the information we need.
Thanks~
machinaut
on 28 Nov 2018
in my case installing libglew-dev did the trick!
kalki123123
on 14 May 2019
No such file or directory: 'patchelf': 'patchelf'
download patchelf 0.9（https://nixos.org/releases/patchelf/patchelf-0.9/patchelf-0.9.tar.gz），unzip
then run ./configure, make, sudo make install，patchelf --version to check if it's ok
PolarisYxh
on 3 Aug 2019
For me a combination of these commands were able to fix this issue:
sudo apt-get install libglew-dev
sudo apt-get install patchelf
sudo ln -s /usr/lib/x86_64-linux-gnu/libGL.so.1 /usr/lib/x86_64-linux-gnu/libGL.so
export LD_PRELOAD=/usr/lib/x86_64-linux-gnu/libGLEW.so
karanbirchahal-bc
on 21 Jun 2020
@karanbirchahal-bc what does the last two commands does?
seyeeet
on 26 Aug 2020
Related issues
JulianYG
·
19Comments
amyzhang
·
25Comments
tadashiK
·
18Comments
marctheshark3
·
13Comments
soerendip
·
22Comments
Most helpful comment
running
sudo apt-get install libglew-devand installing patchelf. I was able to install mujoco v1.5.