Unable to install tensor-flow in R-Studio on m1 mac

[pmgarafola]

Hello - I can't seem to get tensor-flow working on my m1 mac. I wonder if it's a pathing problem or I have the tensor-flow mac-os library installed in the wrong place?

Thanks in advance for your help.

Mac Version: macOS Monterey, version 12.3.1

This is my R-version:

version
_
platform aarch64-apple-darwin20
arch aarch64
os darwin20
system aarch64, darwin20
status
major 4
minor 2.0
year 2022
month 04
day 22
svn rev 82229
language R
version.string R version 4.2.0 (2022-04-22)
nickname Vigorous Calisthenics

Performed the follwing steps in R-studio R script:

install.packages("remotes")
remotes::install_github(sprintf("rstudio/%s", c("reticulate", "tensorflow", "keras")))
reticulate::miniconda_uninstall()

start with a blank slate

reticulate::install_miniconda(force=TRUE)
keras::install_keras()

Everything worked up until the last step. Here is the error message:

keras::install_keras()
Installing: pandas, Pillow, pydot

• '/Users/petergarafola/Library/r-miniconda-arm64/bin/conda' 'install' '--yes' '--prefix' '/usr/local/bin' '-c' 'conda-forge' 'pandas' 'Pillow' 'pydot'
  Collecting package metadata (current_repodata.json): ...working... done
  Solving environment: ...working... done

All requested packages already installed.

Installing: tensorflow-deps

• '/Users/petergarafola/Library/r-miniconda-arm64/bin/conda' 'install' '--yes' '--prefix' '/usr/local/bin' '-c' 'apple' '-c' 'conda-forge' 'tensorflow-deps'
  Collecting package metadata (current_repodata.json): ...working... done
  Solving environment: ...working... done

All requested packages already installed.

Installing: tensorflow-macos

• . /Users/petergarafola/Library/r-miniconda-arm64/bin/activate
• conda activate '/usr/local/bin'
• '/usr/local/bin/bin/python' -m pip install --upgrade 'tensorflow-macos'
  ERROR: Could not find a version that satisfies the requirement tensorflow-macos (from versions: none)
  ERROR: No matching distribution found for tensorflow-macos
  Error: Error installing package(s): "'tensorflow-macos'"

Path environment variable:
/Users/petergarafola/miniforge3/bin:/Users/petergarafola/miniforge3/condabin:/opt/homebrew/bin:/opt/homebrew/sbin:/Library/Frameworks/Python.framework/Versions/3.10/bin:/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin:/Library/Apple/usr/bin

[pmgarafola]

Update - still not working but closer.

I repeated the install steps from scratch. Went into R- Studio and selected R Studio->Preferences->Python. Picked the Python 3.8.9 [System Interpreter] option. Then repeated the tensorflow::install_tensorflow() command. This time it seemed to work and did not give any error messages.

However after loading the keras reticulate, and tensorflow libraries, and then making a call to keras_model_sequential, I get the following:

modelLr = keras_model_sequential()
List of 20
$ python : chr "/usr/bin/python3"
$ libpython : chr "/Library/Developer/CommandLineTools/Library/Frameworks/Python3.framework/Versions/3.8/lib/python3.8/config-3.8-"| truncated
$ pythonhome : chr "/Library/Developer/CommandLineTools/Library/Frameworks/Python3.framework/Versions/3.8:/Library/Developer/Comman"| truncated
$ pythonpath : chr "/Library/Developer/CommandLineTools/Library/Frameworks/Python3.framework/Versions/3.8/lib/python38.zip:/Library"| truncated
$ prefix : chr "/Library/Developer/CommandLineTools/Library/Frameworks/Python3.framework/Versions/3.8"
$ exec_prefix : chr "/Library/Developer/CommandLineTools/Library/Frameworks/Python3.framework/Versions/3.8"
$ base_exec_prefix : chr "/Library/Developer/CommandLineTools/Library/Frameworks/Python3.framework/Versions/3.8"
$ virtualenv : chr ""
$ virtualenv_activate : chr ""
$ version_string : chr "3.8.9 (default, Mar 30 2022, 13:51:16) [Clang 13.1.6 (clang-1316.0.21.2.3)]"
$ version : chr "3.8"
$ architecture : chr "64bit"
$ anaconda : logi FALSE
$ conda : logi FALSE
$ numpy : NULL
$ required_module : chr "tensorflow"
$ required_module_path: NULL
$ available : logi TRUE
$ python_versions : chr "/usr/bin/python3"
$ forced : chr "use_python function"

• attr(*, "class")= chr "py_config"
  Error: Python module tensorflow.keras was not found.

Detected Python configuration:

[njtierney]

Wanted to add that I cannot install Tensorflow on an M1 mac at the moment, I'm not sure what has changed, as I used to be able to do this with no issues.

Here's my reprex + session info:

library(tensorflow)
install_tensorflow(envname = "greta-env")
#> Installing: tensorflow-deps==2.6
#> Error: one or more Python packages failed to install [error code 1]

^{Created on 2022-05-11 by the reprex package (v2.0.1)}

Session info

sessioninfo::session_info()
#> ─ Session info ───────────────────────────────────────────────────────────────
#>  setting  value
#>  version  R version 4.2.0 (2022-04-22)
#>  os       macOS Big Sur 11.2.2
#>  system   aarch64, darwin20
#>  ui       X11
#>  language (EN)
#>  collate  en_AU.UTF-8
#>  ctype    en_AU.UTF-8
#>  tz       Australia/Perth
#>  date     2022-05-11
#>  pandoc   2.17.1.1 @ /Applications/RStudio.app/Contents/MacOS/quarto/bin/ (via rmarkdown)
#> 
#> ─ Packages ───────────────────────────────────────────────────────────────────
#>  package     * version    date (UTC) lib source
#>  base64enc     0.1-3      2015-07-28 [1] CRAN (R 4.2.0)
#>  cli           3.3.0      2022-04-25 [1] CRAN (R 4.2.0)
#>  digest        0.6.29     2021-12-01 [1] CRAN (R 4.2.0)
#>  evaluate      0.15       2022-02-18 [1] CRAN (R 4.2.0)
#>  fastmap       1.1.0      2021-01-25 [1] CRAN (R 4.2.0)
#>  fs            1.5.2      2021-12-08 [1] CRAN (R 4.2.0)
#>  glue          1.6.2      2022-02-24 [1] CRAN (R 4.2.0)
#>  highr         0.9        2021-04-16 [1] CRAN (R 4.2.0)
#>  htmltools     0.5.2      2021-08-25 [1] CRAN (R 4.2.0)
#>  jsonlite      1.8.0      2022-02-22 [1] CRAN (R 4.2.0)
#>  knitr         1.39       2022-04-26 [1] CRAN (R 4.2.0)
#>  lattice       0.20-45    2021-09-22 [1] CRAN (R 4.2.0)
#>  magrittr      2.0.3      2022-03-30 [1] CRAN (R 4.2.0)
#>  Matrix        1.4-1      2022-03-23 [1] CRAN (R 4.2.0)
#>  png           0.1-7      2013-12-03 [1] CRAN (R 4.2.0)
#>  Rcpp          1.0.8.3    2022-03-17 [1] CRAN (R 4.2.0)
#>  reprex        2.0.1      2021-08-05 [1] CRAN (R 4.2.0)
#>  reticulate    1.24       2022-01-26 [1] CRAN (R 4.2.0)
#>  rlang         1.0.2      2022-03-04 [1] CRAN (R 4.2.0)
#>  rmarkdown     2.14       2022-04-25 [1] CRAN (R 4.2.0)
#>  rstudioapi    0.13       2020-11-12 [1] CRAN (R 4.2.0)
#>  sessioninfo   1.2.2      2021-12-06 [1] CRAN (R 4.2.0)
#>  stringi       1.7.6      2021-11-29 [1] CRAN (R 4.2.0)
#>  stringr       1.4.0      2019-02-10 [1] CRAN (R 4.2.0)
#>  tensorflow  * 2.8.0.9000 2022-05-11 [1] Github (rstudio/tensorflow@6e3e00e)
#>  tfruns        1.5.0      2021-02-26 [1] CRAN (R 4.2.0)
#>  whisker       0.4        2019-08-28 [1] CRAN (R 4.2.0)
#>  withr         2.5.0      2022-03-03 [1] CRAN (R 4.2.0)
#>  xfun          0.31       2022-05-10 [1] CRAN (R 4.2.0)
#>  yaml          2.3.5      2022-02-21 [1] CRAN (R 4.2.0)
#> 
#>  [1] /Library/Frameworks/R.framework/Versions/4.2-arm64/Resources/library
#> 
#> ──────────────────────────────────────────────────────────────────────────────

Not all of the error message was captured above, so here is what I get when I run the above code:

> tensorflow::install_tensorflow(envname = "greta-env")
Installing: tensorflow-deps==2.6
Collecting package metadata (current_repodata.json): ...working... done
Solving environment: ...working... failed with initial frozen solve. Retrying with flexible solve.
Solving environment: ...working... failed with repodata from current_repodata.json, will retry with next repodata source.
Collecting package metadata (repodata.json): ...working... done
Solving environment: ...working... failed with initial frozen solve. Retrying with flexible solve.
                                                                                           
Found conflicts! Looking for incompatible packages.
This can take several minutes.  Press CTRL-C to abort.
failed

UnsatisfiableError: The following specifications were found to be incompatible with each other:

Output in format: Requested package -> Available versions

Error: one or more Python packages failed to install [error code 1]

I've tried removing miniconda, and the greta-env environment variable, and I still don't get any progress than this.

[t-kalinowski]

That's unfortunate, looks like an update to a conda package is causing the installation of a pinned tensorflow-macos to fail. I unfortunately don't have an older version of macOS to test on. Can you possibly upgrade to Monterey? Everything should work there.

[t-kalinowski]

Quick update, I confirmed that this successfully installs TensorFlow on an M1 Mac running Monterey

remotes::install_github("rstudio/tensorflow")
reticulate::miniconda_uninstall()
reticulate::install_miniconda()
tensorflow::install_tensorflow()

[pmgarafola]

I uninstalled the tensorflow and keras packages in R Studio and then re-installed tensorflow using the above commands. This works (i.e. keras and tensorflow library calls work) as long as there is nothing selected in the R Studio->Preferences->Python option. As soon as I select one of the available interpreters I get the error message below about tensorflow not being installed, even though it is definitely installed. It must be either a pathing issue or maybe I installed tensorflow wrong. I am hoping that this does not mean I will have similar issues with other python libraries, but I can use the tensorflow library for now. Thanks for the help.

Error: Valid installation of TensorFlow not found.

Python environments searched for 'tensorflow' package:
/Library/Frameworks/Python.framework/Versions/3.10/bin/python3.10

Python exception encountered:
Traceback (most recent call last):
File "/Library/Frameworks/R.framework/Versions/4.2-arm64/Resources/library/reticulate/python/rpytools/loader.py", line 39, in _import_hook
module = _import(
ModuleNotFoundError: No module named 'tensorflow'

You can install TensorFlow using the install_tensorflow() function.

[t-kalinowski]

Hi, I think it's working as intended. tensorflow::install_tensorflow() installs the Python 'tensorflow' package into a single specific Python installation. Any given computer may have multiple Python installations, and if you want to use Tensorflow with some other Python installation (not the default "r-reticulate" conda environment in the r-miniconda Python installation reticulate bootstraps), you'll have to point tensorflow::install_tensorflow() at it, e.g, by passing a valid value to envname.

[pmgarafola]

How does the install select a python installation if you don't specify an environment? Or are you saying that the install creates a python installation for tensorflow to use?

As I mentioned, I have it working so am closing the issue. Thanks for the help.

[t-kalinowski]

install_miniconda() creates a python installation ("r-miniconda" under rappdirs::user_data_dir()), and also creates a conda environment with that installation named r-reticulate. Subsequently, reticulate::py_install() selects the r-reticulate conda environment as the default, if you've taken no explicit action opting to use another Python.