【发布时间】：2022-02-03 01:17:57
【问题描述】：
lazypredict 需要 numpy==1.19.1 版本。但是，这个版本的 numpy 与 tensorflow 不兼容。我尝试安装 tensorflow 2.2.0、2.3.0、2.4.0 和 2.5.0，但这些都不兼容 numpy==1.19.1。
我可以使用哪个版本的 tensorflow 来避免延迟预测所需的包的依赖问题？
【问题讨论】：
标签： python numpy tensorflow
【发布时间】：2022-02-03 01:17:57
【问题描述】：
令人讨厌的是，由于 setup.py 文件 (https://dustingram.com/articles/2018/03/05/why-pypi-doesnt-know-dependencies/) 的潜在不确定行为，pypi 无法真正列出包的依赖关系。
但是，当我尝试使用 anaconda 安装这两个软件包时：conda install lazypredict
过了一会儿，它安装了lazypredict-0.2.9和tensorflow-2.4.3
老实说，如果简单地安装两个软件包的最新版本而忽略不匹配的依赖项会导致任何问题，我会感到非常惊讶（因为 numpy 通常向后兼容）。
编辑：现在我尝试从上面安装提到的版本，但 pip 仍然抱怨：tensorflow 2.4.3 requires numpy~=1.19.2, but you have numpy 1.19.1 which is incompatible. 因此，我添加了 anaconda 想要安装的软件包的完整列表。我不确定~= 到底是什么意思，但显然 anaconda 认为 1.19.1 满足该要求。
The following packages will be downloaded:
package | build
---------------------------|-----------------
_py-xgboost-mutex-2.0 | cpu_0 8 KB conda-forge
abseil-cpp-20210324.2 | he49afe7_0 937 KB conda-forge
absl-py-1.0.0 | pyhd8ed1ab_0 95 KB conda-forge
aiohttp-3.8.1 | py37h271585c_0 562 KB conda-forge
aiosignal-1.2.0 | pyhd8ed1ab_0 12 KB conda-forge
astor-0.8.1 | pyh9f0ad1d_0 25 KB conda-forge
astunparse-1.6.3 | pyhd8ed1ab_0 15 KB conda-forge
async-timeout-4.0.2 | pyhd8ed1ab_0 9 KB conda-forge
asynctest-0.13.0 | py_0 24 KB conda-forge
attrs-21.4.0 | pyhd8ed1ab_0 49 KB conda-forge
backports-1.0 | py_2 4 KB conda-forge
backports.functools_lru_cache-1.6.4| pyhd8ed1ab_0 9 KB conda-forge
blinker-1.4 | py_1 13 KB conda-forge
brotlipy-0.7.0 |py37h271585c_1003 357 KB conda-forge
c-ares-1.18.1 | h0d85af4_0 100 KB conda-forge
ca-certificates-2021.10.8 | h033912b_0 139 KB conda-forge
cachetools-4.2.4 | pyhd8ed1ab_0 12 KB conda-forge
certifi-2021.10.8 | py37hf985489_1 145 KB conda-forge
cffi-1.14.6 | py37hda351d9_0 218 KB conda-forge
charset-normalizer-2.0.11 | pyhd8ed1ab_0 35 KB conda-forge
click-7.1.2 | pyh9f0ad1d_0 64 KB conda-forge
cryptography-36.0.1 | py37h5e77fcc_0 1.3 MB conda-forge
dataclasses-0.8 | pyhc8e2a94_3 10 KB conda-forge
frozenlist-1.3.0 | py37h271585c_0 39 KB conda-forge
gast-0.3.3 | py_0 12 KB conda-forge
giflib-5.2.1 | hbcb3906_2 71 KB conda-forge
google-auth-1.35.0 | pyh6c4a22f_0 81 KB conda-forge
google-auth-oauthlib-0.4.6 | pyhd8ed1ab_0 19 KB conda-forge
google-pasta-0.2.0 | pyh8c360ce_0 42 KB conda-forge
grpc-cpp-1.37.1 | h69eed7d_3 2.8 MB conda-forge
grpcio-1.37.1 | py37h87f55f7_0 1.8 MB conda-forge
h5py-2.10.0 |nompi_py37hdf859c4_106 905 KB conda-forge
hdf5-1.10.6 |nompi_hc5d9132_1114 3.1 MB conda-forge
icu-68.2 | he49afe7_0 12.7 MB conda-forge
idna-3.3 | pyhd8ed1ab_0 55 KB conda-forge
importlib-metadata-4.10.1 | py37hf985489_0 32 KB conda-forge
importlib_metadata-4.10.1 | hd8ed1ab_0 4 KB conda-forge
joblib-1.0.0 | pyhd8ed1ab_0 206 KB conda-forge
jpeg-9e | h0d85af4_0 252 KB conda-forge
keras-preprocessing-1.1.2 | pyhd8ed1ab_0 34 KB conda-forge
krb5-1.19.2 | hcfbf3a7_3 1.2 MB conda-forge
lazypredict-0.2.9 | pyhd8ed1ab_0 15 KB conda-forge
libblas-3.9.0 |13_osx64_openblas 12 KB conda-forge
libcblas-3.9.0 |13_osx64_openblas 12 KB conda-forge
libcurl-7.79.1 | hf45b732_1 317 KB conda-forge
libedit-3.1.20191231 | h0678c8f_2 103 KB conda-forge
libev-4.33 | haf1e3a3_1 99 KB conda-forge
libgfortran-5.0.0 |9_3_0_h6c81a4c_23 19 KB conda-forge
libgfortran5-9.3.0 | h6c81a4c_23 1.7 MB conda-forge
liblapack-3.9.0 |13_osx64_openblas 12 KB conda-forge
libnghttp2-1.43.0 | h6f36284_1 867 KB conda-forge
libopenblas-0.3.18 |openmp_h3351f45_0 9.0 MB conda-forge
libpng-1.6.37 | h7cec526_2 313 KB conda-forge
libprotobuf-3.15.8 | hcf210ce_1 2.2 MB conda-forge
libssh2-1.10.0 | h52ee1ee_2 221 KB conda-forge
libxgboost-1.5.1 | h4a89273_0 2.2 MB conda-forge
lightgbm-2.3.1 | py37h4a8c4bd_0 715 KB conda-forge
llvm-openmp-12.0.1 | hda6cdc1_1 287 KB conda-forge
markdown-3.3.6 | pyhd8ed1ab_0 67 KB conda-forge
more-itertools-8.12.0 | pyhd8ed1ab_0 47 KB conda-forge
multidict-6.0.2 | py37h271585c_0 45 KB conda-forge
numpy-1.19.1 | py37h1efc2f6_2 5.1 MB conda-forge
oauthlib-3.2.0 | pyhd8ed1ab_0 90 KB conda-forge
openssl-1.1.1l | h0d85af4_0 1.9 MB conda-forge
opt_einsum-3.3.0 | pyhd8ed1ab_1 53 KB conda-forge
packaging-21.3 | pyhd8ed1ab_0 36 KB conda-forge
pandas-1.0.5 | py37h94625e5_0 9.6 MB conda-forge
pluggy-0.13.1 | py37hf985489_4 29 KB conda-forge
protobuf-3.15.8 | py37hd8d24ac_0 323 KB conda-forge
py-1.11.0 | pyh6c4a22f_0 74 KB conda-forge
py-xgboost-1.5.1 | py37hf985489_0 150 KB conda-forge
pyasn1-0.4.8 | py_0 53 KB conda-forge
pyasn1-modules-0.2.7 | py_0 60 KB conda-forge
pycparser-2.21 | pyhd8ed1ab_0 100 KB conda-forge
pyjwt-2.3.0 | pyhd8ed1ab_1 18 KB conda-forge
pyopenssl-22.0.0 | pyhd8ed1ab_0 49 KB conda-forge
pyparsing-3.0.7 | pyhd8ed1ab_0 79 KB conda-forge
pysocks-1.7.1 | py37hf985489_4 28 KB conda-forge
pytest-5.4.3 | py37hc8dfbb8_0 384 KB conda-forge
python-dateutil-2.8.2 | pyhd8ed1ab_0 240 KB conda-forge
python-flatbuffers-1.12 | pyhd8ed1ab_1 19 KB conda-forge
python_abi-3.7 | 2_cp37m 4 KB conda-forge
pytz-2021.3 | pyhd8ed1ab_0 242 KB conda-forge
pyu2f-0.1.5 | pyhd8ed1ab_0 31 KB conda-forge
pyyaml-5.3.1 | py37hf967b71_2 171 KB conda-forge
re2-2021.08.01 | he49afe7_0 193 KB conda-forge
requests-2.27.1 | pyhd8ed1ab_0 53 KB conda-forge
requests-oauthlib-1.3.1 | pyhd8ed1ab_0 22 KB conda-forge
rsa-4.8 | pyhd8ed1ab_0 31 KB conda-forge
scikit-learn-0.23.1 | py37hf5857e7_0 6.3 MB conda-forge
scipy-1.7.3 | py37h4e3cf02_0 19.7 MB conda-forge
six-1.15.0 | pyh9f0ad1d_0 14 KB conda-forge
snappy-1.1.8 | hb1e8313_3 29 KB conda-forge
tensorboard-2.4.1 | pyhd8ed1ab_1 8.8 MB conda-forge
tensorboard-plugin-wit-1.8.1| pyhd8ed1ab_0 668 KB conda-forge
tensorflow-2.4.3 | py37hf985489_0 25 KB conda-forge
tensorflow-base-2.4.3 | py37h05ab19a_0 116.3 MB conda-forge
tensorflow-estimator-2.4.0 | pyh9656e83_0 289 KB conda-forge
termcolor-1.1.0 | py_2 6 KB conda-forge
threadpoolctl-3.1.0 | pyh8a188c0_0 18 KB conda-forge
tqdm-4.56.0 | pyhd8ed1ab_0 76 KB conda-forge
typing-extensions-3.7.4.3 | 0 8 KB conda-forge
typing_extensions-3.7.4.3 | py_0 25 KB conda-forge
urllib3-1.26.8 | pyhd8ed1ab_1 100 KB conda-forge
wcwidth-0.2.5 | pyh9f0ad1d_2 33 KB conda-forge
werkzeug-2.0.2 | pyhd8ed1ab_0 221 KB conda-forge
wrapt-1.13.3 | py37h271585c_1 44 KB conda-forge
xgboost-1.5.1 | py37h3f6e5c9_0 12 KB conda-forge
yaml-0.2.5 | h0d85af4_2 82 KB conda-forge
yarl-1.7.2 | py37h271585c_1 127 KB conda-forge
zipp-3.7.0 | pyhd8ed1ab_0 12 KB conda-forge
------------------------------------------------------------
Total: 217.4 MB
【讨论】：
-
~=1.19.2表示，选择至少为 1.19.2 的版本。它将支持更大的版本，但在 1.19.* 系列中，而不是 1.20.*。这个解释更详细here