Installing farm-haystack causes weird errors

[Obsideaock]

Hi, I'm having a strange problem when installing farm-haystack with either
pip install farm-haystack -f https://download.pytorch.org/whl/torch_stable.html --user
or
pip install farm-haystack

It appears everything correctly installs, but then this error appears in the terminal
ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behavior is the source of the following dependency conflicts. farm 0.8.0 requires transformers==4.6.1, but you have transformers 4.7.0 which is incompatible. datasets 1.15.2.dev0 requires huggingface-hub<1.0.0,>=0.1.0, but you have huggingface-hub 0.0.8 which is incompatible.

I've tried directly install the packages with the versions it wants, but that's only lead to more incompatibilities. Can anyone please help?

[bogdankostic]

Hi @Obsideaock! I was not able to replicate your issue on my MacOS machine. Which operating system are you using? Did you try to install haystack inside a new environment?

[superchargez]

Hi, I'm having a strange problem when installing farm-haystack with either pip install farm-haystack -f https://download.pytorch.org/whl/torch_stable.html --user or pip install farm-haystack

It appears everything correctly installs, but then this error appears in the terminal ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behavior is the source of the following dependency conflicts. farm 0.8.0 requires transformers==4.6.1, but you have transformers 4.7.0 which is incompatible. datasets 1.15.2.dev0 requires huggingface-hub<1.0.0,>=0.1.0, but you have huggingface-hub 0.0.8 which is incompatible.

I've tried directly install the packages with the versions it wants, but that's only lead to more incompatibilities. Can anyone please help?

Upgrade pip. I had same issue for another framework but upgrading pip resolved this issue.
You may encounter problems upgrading pip (depending on 1> platform 2> already conda based environment)
1a> If you are using windows just reinstall pip/python (after deleting older version, this is the fastest option)
1b> If you are on linux install pyenv and install python that you need and THEN install/upgrade pip
1b2> if above does not work then remove pyenv and install conda (I mean switch environment builder to another) and upgrade pip
2> for conda environment users, create a new environment and just upgrade pip