Hello
I am trying nsynth training, but an error occurs.If there is a solution, please let me know
The environment is
Ubuntu 16.04 LTS
magenta-gpu 0.3.12
tensorflow-gpu1.11.0
$bazel run//magenta/models/nsynth/baseline:train ----train_path=//home/nekome/tffile.tfrecords---logdir=//home/nekome/Logdire
WARNING: Processed legacy workspace file/home/nekome/magenta/tools/bazel.rc.This file will not be processed in the next release of Bazel.Please read https://github.com/bazelbuild/bazel/issues/6319 for further information, including how to upgrade.
INFO: Analyzed target // magenta/models/nsynth/baseline: train (0 packages loaded).
INFO: Found 1 target...
Target // magenta/models/nsynth/baseline:train up-to-date:
bazel-bin/magenta/models/nsynth/baseline/train
INFO: Elapped time: 0.059s, Critical Path: 0.00s
INFO: 0 processes.
INFO—Build completed successfully, 1 total action
INFO:running command line:bazel-bin/magenta/models/nsynth/baseline/train' --trINFO:Build completed successfully, 1 total action
Itensorflow/stream_executor/dso_loader.cc:105] successfully opened CUDA library libcublas.so locally
Itensorflow/stream_executor/dso_loader.cc:99]Couldn't open CUDA library libcudnn.so.LD_LIBRARY_PATH:
Itensorflow/stream_executor/cuda/cuda_dnn.cc:1407] Unable to load cuDNN DSO
Itensorflow/stream_executor/dso_loader.cc:105] successfully opened CUDA library libcufft.so locally
Itensorflow/stream_executor/dso_loader.cc:105] successfully opened CUDA library libcuda.so.1 locally
Itensorflow/stream_executor/dso_loader.cc:105] successfully opened CUDA library libcurand.so locally
Traceback (most recent call last):
File"/home/nekome/.cache/bazel/_bazel_root/8d760b4887a3a97dbc093bfa28201502/execroot/_main__/bazel-out/k8-opt/bin/magenta/models/nsynth/baseline/train.runfiles/_main__/main_magency/line, 23 train.py
from magenta.models.nsynth import reader
File"/home/nekome/.cache/bazel/_bazel_root/8d760b4887a3a97dbc093bfa28201502/execroot/_main__/bazel-out/k8-opt/bin/magenta/models/nsynth/baseline/train.runfiles/_main__/models, reader.py
from magenta.models.nsynth import utils
File"/home/nekome/.cache/bazel/_bazel_root/8d760b4887a3a97dbc093bfa28201502/execroot/_main__/bazel-out/k8-opt/bin/magenta/models/nsynth/baseline/train.runfiles/_main__/models, utils.py
slim = tf.contrib.slim
AttributeError: 'module' object has no attribute' slim'
How can I solve this problem?
The file utils.py in question is here
This question appears to be about building a Magenta development environment.If you encounter an error like the question, it is highly likely that the package is experiencing problems, so you should create a new virtual environment and follow the instructions on the Magenta home page (https://github.com/tensorflow/magenta).Install in minutes.
Ubuntu 16.04 LTS to use Python 3:
If venv has not been successfully installed, install it with apt.
sudo apt install python3-venv
Then, navigate to the directory where you want to install the Magenta development environment.
Next, Coulomb the repository of agmagenta` and navigate to that directory.
git clone https://github.com/tensorflow/magenta.git
cd./magenta
Create a virtual environment
/usr/bin/python3-mvenvenv
After enabling the virtual environment, install magenta-gpu. Installing magenta-gpu with pip also installs other required packages such as tensorflow.
source env/bin/activate
pip3 install magenta-gpu
Test the Magenta development environment.
bazel test//magenta/...
Executed 1 out of 55 tests:54 tests pass and 1 failures locally. is OK. // magenta/music:musicnet_io_test fails for Python 3.
Run the following command when you exit the virtual environment:
deactivate
From now on, go to the Magenta development environment directory and enable the virtual environment below before you work.
source env/bin/activate
For a more detailed description of the virtual environment, see 12. Virtual environment and package in the official tutorial.You can also create a virtual environment if you want to use Python 2, Anaconda (Miniconda).If so, check your documentation.
© 2024 OneMinuteCode. All rights reserved.