Anaconda3を使って、研究室用に機械学習を学ぶ環境を構築していた時のこと。
tensorflowのインストール・importができず3時間費やした。
結論
m1Macにはm1Mac用のtensorflowを使う必要があった
これに気づかず、時間がかかってしまった
解決までの過程
試したがうまくいかなかったこと
$ pip install tensorflow
ERROR: Could not find a version that satisfies the requirement tensorflow==2.7.0 (from versions: none)
ERROR: No matching distribution found for tensorflow==2.7.0
pipのアップグレード
pip install --upgrade pip
anacondaからpythonのバージョンを3.10から３.9や３.8に下げてみるができない
以下コマンドでインストールはできた
pip3 install --upgrade --ignore-installed https://storage.googleapis.com/tensorflow/mac/cpu/tensorflow-1.12.0-py3-none-any.whl
↓参考にした
importできない問題
installできた！と思ったら、今度はimportができない状況に
import tensorflow
ImportError Traceback (most recent call last)
File ~/anaconda3/envs/study/lib/python3.9/site-pacages/tensorflow/python/pywrap_tensorflow.py:58
56 sys.setdlopenflags(_default_dlopen_flags | ctypes.RTLD_LOCAL)
---> 58 from tensorflow.python.pywrap_tensorflow_internal import *
59 from tensorflow.python.pywrap_tensorflow_internal import __version__
File ~/anaconda3/envs/study/lib/python3.9/site-packages/tensorflow/python/pywrap_tensorflow_internal.py:28
27 return _mod
---> 28 _pywrap_tensorflow_internal = swig_import_helper()
29 del swig_import_helper
File ~/anaconda3/envs/study/lib/python3.9/site-packages/tensorflow/python/pywrap_tensorflow_internal.py:24, in swig_import_helper()
23 try:
---> 24 _mod = imp.load_module('_pywrap_tensorflow_internal', fp, pathname, description)
25 finally:
File ~/anaconda3/envs/study/lib/python3.9/imp.py:242, in load_module(name, file, filename, details)
241 else:
--> 242 return load_dynamic(name, filename, file)
243 elif type_ == PKG_DIRECTORY:
File ~/anaconda3/envs/study/lib/python3.9/imp.py:342, in load_dynamic(name, path, file)
340 spec = importlib.machinery.ModuleSpec(
341 name=name, loader=loader, origin=path)
--> 342 return _load(spec)
ImportError: dlopen(/Users/username/anaconda3/envs/study/lib/python3.9/site-packages/tensorflow/python/_pywrap_tensorflow_internal.so, 0x0006): tried: '/Users/username/anaconda3/envs/study/lib/python3.9/site-packages/tensorflow/python/_pywrap_tensorflow_internal.so' (mach-o file, but is an incompatible architecture (have 'x86_64', need 'arm64')), '/System/Volumes/Preboot/Cryptexes/OS/Users/username/anaconda3/envs/study/lib/python3.9/site-packages/tensorflow/python/_pywrap_tensorflow_internal.so' (no such file), '/Users/username/anaconda3/envs/study/lib/python3.9/site-packages/tensorflow/python/_pywrap_tensorflow_internal.so' (mach-o file, but is an incompatible architecture (have 'x86_64', need 'arm64'))
During handling of the above exception, another exception occurred:
ImportError Traceback (most recent call last)
Cell In[2], line 1
----> 1 import tensorflow
File ~/anaconda3/envs/study/lib/python3.9/site-packages/tensorflow/__init__.py:24
21 import os as _os
23 # pylint: disable=g-bad-import-order
---> 24 from tensorflow.python import pywrap_tensorflow # pylint: disable=unused-import
26 try:
27 # Add `estimator` attribute to allow access to estimator APIs via
28 # "tf.estimator..."
29 from tensorflow.python.estimator.api import estimator # pylint: disable=g-import-not-at-top
File ~/anaconda3/envs/study/lib/python3.9/site-packages/tensorflow/python/__init__.py:49
32 # TODO(drpng): write up instructions for editing this file in a doc and point to
33 # the doc instead.
34 # If you want to edit this file to expose modules in public tensorflow API, you
(...)
44 # go/tf-wildcard-import
45 # pylint: disable=wildcard-import,g-bad-import-order,g-import-not-at-top
47 import numpy as np
---> 49 from tensorflow.python import pywrap_tensorflow
51 from tensorflow.python.tools import component_api_helper
52 component_api_helper.package_hook(
53 parent_package_str='tensorflow.python',
54 child_package_str=(
55 'tensorflow_estimator.python.estimator'))
File ~/anaconda3/envs/study/lib/python3.9/site-packages/tensorflow/python/pywrap_tensorflow.py:74
69 except ImportError:
70 msg = """%s\n\nFailed to load the native TensorFlow runtime.\n
71 See https://www.tensorflow.org/install/errors\n
72 for some common reasons and solutions. Include the entire stack trace
73 above this error message when asking for help.""" % traceback.format_exc()
---> 74 raise ImportError(msg)
ImportError: Traceback (most recent call last):
File "/Users/username/anaconda3/envs/study/lib/python3.9/site-packages/tensorflow/python/pywrap_tensorflow.py", line 58, in <module>
from tensorflow.python.pywrap_tensorflow_internal import *
File "/Users/username/anaconda3/envs/study/lib/python3.9/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 28, in <module>
_pywrap_tensorflow_internal = swig_import_helper()
File "/Users/username/anaconda3/envs/study/lib/python3.9/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 24, in swig_import_helper
_mod = imp.load_module('_pywrap_tensorflow_internal', fp, pathname, description)
File "/Users/username/anaconda3/envs/study/lib/python3.9/imp.py", line 242, in load_module
return load_dynamic(name, filename, file)
File "/Users/username/anaconda3/envs/study/lib/python3.9/imp.py", line 342, in load_dynamic
return _load(spec)
ImportError: dlopen(/Users/username/anaconda3/envs/study/lib/python3.9/site-packages/tensorflow/python/_pywrap_tensorflow_internal.so, 0x0006): tried: '/Users/username/anaconda3/envs/study/lib/python3.9/site-packages/tensorflow/python/_pywrap_tensorflow_internal.so' (mach-o file, but is an incompatible architecture (have 'x86_64', need 'arm64')), '/System/Volumes/Preboot/Cryptexes/OS/Users/username/anaconda3/envs/study/lib/python3.9/site-packages/tensorflow/python/_pywrap_tensorflow_internal.so' (no such file), '/Users/username/anaconda3/envs/study/lib/python3.9/site-packages/tensorflow/python/_pywrap_tensorflow_internal.so' (mach-o file, but is an incompatible architecture (have 'x86_64', need 'arm64'))
Failed to load the native TensorFlow runtime.
See https://www.tensorflow.org/install/errors
for some common reasons and solutions. Include the entire stack trace
above this error message when asking for help.
ここでm1macにはm1mac用のtensorflow-macosを使う必要があることを知る
$ conda install -c apple tensorflow-deps==2.8.0
$ python -m pip install tensorflow-macos==2.8.0
$ python -m pip install tensorflow-metal
これで解決と思いきや、
TypeError: Descriptors cannot not be created directly.
If this call came from a _pb2.py file, your generated code is out of date and must be regenerated with protoc >= 3.19.0.
If you cannot immediately regenerate your protos, some other possible workarounds are:
1. Downgrade the protobuf package to 3.20.x or lower.
2. Set PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION=python (but this will use pure-Python parsing and will be much slower).
protbuf（Googleが出したシリアライズフォーマット）のバージョンが高いことが原因だったみたいです
下げてあげます
(study) / ᐅ pip install protobuf==3.20.1
Collecting protobuf==3.20.1
Downloading protobuf-3.20.1-py2.py3-none-any.whl (162 kB)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 162.1/162.1 kB 2.7 MB/s eta 0:00:00
Installing collected packages: protobuf
Attempting uninstall: protobuf
Found existing installation: protobuf 4.22.4
Uninstalling protobuf-4.22.4:
Successfully uninstalled protobuf-4.22.4
importできるようになりました！
参考記事