sopimlab / deeplearningwithaudio Goto Github PK

Great initiative, thanks.
Unfortunately flext does not build all targets, which affects py and subsequently the whole build stalls.

OS 10.14.6
Xcode 11.3.1
pd-0.50.2

The error
`/Applications/Xcode.app/Contents/Developer/usr/bin/make -f /Users/cerkut/Developer/DeepLearningWithAudio/pyext-setup/flext/buildsys/gnumake-sub.mak PLATFORM=mac RTSYS=pd COMPILER=gcc BUILDPATH=/Users/cerkut/Developer/DeepLearningWithAudio/pyext-setup/flext/buildsys/ PKGINFO=package.txt BUILDCLASS=flext USRCONFIG=config.txt USRMAKE=build/gnumake-mac-gcc.inc TARGETMODE=debug TARGETTYPE=single DEBUG=1 install

/Users/cerkut/Developer/DeepLearningWithAudio/pyext-setup/flext/buildsys/mac/gnumake-gcc-targets.inc:22: warning: overriding commands for target `pd-darwin/debug-single'

/Users/cerkut/Developer/DeepLearningWithAudio/pyext-setup/flext/buildsys/mac/gnumake-gcc-targets.inc:18: warning: ignoring old commands for target `pd-darwin/debug-single'

install pd-darwin/debug-single/libflext-pd_sd.0.6.1.a /Users/cerkut/Developer/DeepLearningWithAudio/pyext-setup/build/flext/lib
install: pd-darwin/debug-single/libflext-pd_sd.0.6.1.a: No such file or directory
make[1]: *** [install] Error 71
make: *** [install-debug-single] Error 2
Traceback (most recent call last):
File "build.py", line 217, in
flext_build(args=("install",))
File "build.py", line 195, in
flext_build = lambda ok_codes=(0,), args=(): call_with_env(flext_build_cmd + args, flext_build_env, ok_codes)
File "build.py", line 119, in call_with_env
raise Exception("command exited with unexpected code {}".format(code))
Exception: command exited with unexpected code 2`

By selectively changing the flags BUILDMODE and TARGETMODE I am investigating where the error originates from, and will report. But appreciate the help if other people have the same problem.

flext_build_env = { "BUILDMODE": "debug", #"TARGETMODE": "single", "FLEXT_PD_APP": pd, "FLEXT_INSTALL_PATH": package_dir, "FLEXTPREFIX": flext_prefix, "FLEXT_NOATTREDIT": str(int(is_purr_data)) }

I'm following the steps for building the second example, 02_ddsp, and I'm getting stuck in pip install google-colab.
The errors start when it's trying to build pandas. I've uploaded the output which I got with pip install google-colab 2> colab_error.txt
I'm on UbuntuStudio 21.04. I created the Conda environment with conda create -n ddsp python=3.8 tensorflow=2.4.1 tensorflow-probability pandas=0.25
Because using pandas=0.24 was giving me incompatibility errors with Python3.8.
colab_error.txt

Hi, I am following the instructions on this page:

https://github.com/SopiMlab/DeepLearningWithAudio/blob/master/utilities/pyext-setup/setup-linux.md

When I reach installing Magenta into the Conda environment from the current directory using pip, then it becomes a death spiral of dependency resolution.

The pip command starts with tensorflow 4.6 and eventually works its way down to tensorflow 1.13 without ever resolving all the dependencies. It fails on note-seq-0.0.4 with the error "pip._vendor.resolvelib.resolvers.ResolutionTooDeep: 2000000" after running for a few hours.

Did I possibly do something wrong or are the pip config files in need of an update somehow?

Thanks!

Instead of putting the py/ directory to Pd's startup paths and the py lib in Pd's startup libraries, I'd rather use [declare -path py -lib py] which works fine for the [py] default help files, but when I'm trying to open the ai_duet.pd patch, Pd segfaults.
Opening Pd with gdb I get this:

Thread 1 "pd" received signal SIGSEGV, Segmentation fault.
0x00007ffff659e464 in _PyEval_EvalFrameDefault () from /home/alexandros/miniconda3/envs/magenta/lib/libpython3.8.so.1.0

I'm on UbuntuStudio 21.04 with Pd-0.51.4

When running the pip install command from the magenta folder.

I'm running a new MacBook Pro M1, macOS 11.6, Pd 0.51.4.

I'm almost certain that I've followed all of the setup correctly, resulting in a build of pyext that is installed in my Pd externals directory. When I launch Pd I see the messages in the Pd window indicated that pyext (and zexy) has loaded without errors.

magenta has also been installed and seems to be activating properly.

I have downloaded the pretrained models, but when I launch ai_duet.pd the patch will not open in Pd. It will sometimes crash Pd too, though not consistently. I'm not sure where the problem is coming from.

The only differences I'm seeing between my setup and the information in the docs is that I'm running Pd 0.51.4 (instead of 0.51.3) and I'm using Python 3.7.11 (instead of 3.7.10).

I recognise that this may not be enough info to go on, but...any ideas?

When running the make_dataset.py script.

Mac mini (M1, 2020)

Whether I put py from DeepLearningWithAudio in the startup paths and libraries or I use [declare -path py -lib py] I cannot quit Pd. The library loads fine and Pd functions properly, it just won't quit.
Launching Pd with gdb I get this when trying to quit:

Thread 1 "pd" received signal SIGHUP, Hangup.
futex_wait (private=<optimized out>, expected=12, futex_word=0x7ffff693126c <pybase::qucond+76>) at ../sysdeps/nptl/futex-internal.h:146
146     ../sysdeps/nptl/futex-internal.h: No such file or directory.

I'm on UbuntuStudio 21.04 with Pd-0.51.4

Trying to compile everything according to https://github.com/SopiMlab/DeepLearningWithAudio/blob/master/utilities/pyext-setup/setup-linux.md
Everything works fine with conda create -n magenta python=3.8 tensorflow=2.3 untill:

cd ../pyext-setup
python build.py

(BTW, the first line shouldn't be cd ../utilities/pyext-setup?)

I get the following error:

make[1]: Entering directory '/home/alexandros/Documents/DeepLearningWithAudio/utilities/pyext-setup/py'
install pd-linux/debug-multi/py.pd_linux /home/alexandros/Documents/DeepLearningWithAudio/utilities/pyext-setup/build/py
install: cannot stat 'pd-linux/debug-multi/py.pd_linux': No such file or directory
make[1]: *** [/home/alexandros/Documents/DeepLearningWithAudio/utilities/pyext-setup/flext/buildsys/lnx/gnumake-gcc-ext.inc:56: _install_] Error 1
make[1]: Leaving directory '/home/alexandros/Documents/DeepLearningWithAudio/utilities/pyext-setup/py'
make: *** [/home/alexandros/Documents/DeepLearningWithAudio/utilities/pyext-setup/flext/buildsys/targets.inc:103: install-debug-multi] Error 2
Traceback (most recent call last):
  File "build.py", line 376, in <module>
    setup.py_build(args=("install",))
  File "build.py", line 200, in py_build
    return call_with_env(
  File "build.py", line 273, in call_with_env
    raise Exception("command exited with unexpected code {}".format(code))
Exception: command exited with unexpected code 2

When I'm trying to build pyext manually with:

cd ../utilities/pyext-setup/py/
make CPPFLAGS="-I ../flext/source"

I get this:

alexandros@alexandros:~/Documents/DeepLearningWithAudio/utilities/pyext-setup/py$ make CPPFLAGS="-I ../flext/source"
++++ info: using Makefile.pdlibbuilder version 0.6.0
++++ info: using Pd API /usr/include/pd/m_pd.h
++++ info: making target all 
++++ info: making source/bound.o in lib 
g++ -DPD -I "/usr/include/pd" -DUNIX -I ../flext/source -fPIC -fcheck-new -DFLEXT_INLINE -DFLEXT_ATTRIBUTES=1 -Wall -Wextra -Wshadow -Winline -Wstrict-aliasing -O3 -ffast-math -funroll-loops -fomit-frame-pointer -march=core2 -mfpmath=sse -msse -msse2 -msse3 -o source/bound.o -c source/bound.cpp
In file included from source/main.h:12,
                 from source/pybase.h:12,
                 from source/pyext.h:12,
                 from source/bound.cpp:9:
source/pyprefix.h:22:10: fatal error: Python.h: No such file or directory
   22 | #include <Python.h>
      |          ^~~~~~~~~~
compilation terminated.
make: *** [pd-lib-builder//Makefile.pdlibbuilder:955: source/bound.o] Error 1

Both Python3.8 and Python3.9 header files are included in /usr/inlcude/python3.8 and /usr/include/python3.9 respectively.
I'm on UbuntuStudio 21.04. I installed Miniconda3 with Python3.9, the top file from https://docs.conda.io/en/latest/miniconda.html#linux-installers
Don't know if that could be an issue.