I cannot install this library. Can someone show me how?

[Led_Zeppelin]

How in the world does one import this file. I have tried many things and nothing so far has worked. I even Googled the subject and again nothing. So how does one install:

official.nlp

It is used in a program as shown:

from official.nlp import optimization  # to create AdamW optimizer

I just need to know how to install it, so I can import it into my program.

Thnaks fo any help.

Respectfully,

LZ

[ndc85430]

It would help to know what you tried and what the problems were. What does the library's website say to do? Is it not available in PyPI?

[buran]

I think you need to install https://pypi.org/project/tf-models-official/

Depending on what you use, you may need to install more dependencies. look at https://github.com/tensorflow/models/tre...r/official

[Led_Zeppelin]

I tried :

pip install official.nlp
conda insat;; official.nlp

I tried

pip install official
conda install official


I tried pyforest, the general library installer.

None of them worked.

I have not tried lazy_imports(). I will this afternoon.


Respectfully,

LZ

[buran]

Did you open the link PyPI that I shared?

pip install tf-models-official

eventually because of the nlp, you may need to install

pip install tensorflow-text

[Led_Zeppelin]

Thank you. I will give it a try. It just seems a awkward way to install a library.

R,

LZ

[ndc85430]

Why awkward? It's the standard way to do it.

[buran]

It just seems a awkward way to install a library.

What is that you find so awkward? Also, it's obvious you didn't write the code yourself, so I expect there were instructions what and how to install.

[snippsat]

The fast way.
Go to Colab,New Notebook.
Run.

!pip install -q tf-models-official==2.4.0

Example Notebook.

With Anaconda/conda/pip you should create a new environment,look at this Thread.
This need to install a lot of packages look at requirements.txt around 30.
Then is better to make new virtual environment(see link),so it don't crash with what you have tried to install before.