Magenta installation using "pip install magenta" having errors even on fresh environments

I have been attempting to install magenta on anaconda for a few days now, but every time I feel like I'm getting close to achieving an actual functional environment that can run it, I run into errors. Originally I thought the issue might be some conflicting versions, so I created a whole new environment and ran "pip install magenta" to hopefully come up with a working environment. That's when I got these two errors during the install:

ERROR: tensorflow 2.2.0 has requirement scipy==1.4.1; python_version >= "3", but you'll have scipy 1.5.1 which is incompatible.
ERROR: tensor2tensor 1.15.7 has requirement tensorflow-probability==0.7.0, but you'll have tensorflow-probability 0.10.1 which is incompatible.

I figured this might be a simple fix, since it was probably just different versions I needed to install, so I proceeded to install the specified versions that it says it requires in the error message. Then I got a completely different error that was structured the same but was slightly different.

ERROR: apache-beam 2.22.0 has requirement oauth2client<4,>=2.0.1, but you'll have oauth2client 4.1.3 which is incompatible.

After receiving this error I decided it would probably be best if I just started all over and tried this again, so I made another new environment and did it all again. Needless to say, my many attempts have given me nothing in return, let alone a functional environment.

I even took the time to reach out to a friend and asked for his help, and he followed similar steps to me and used the pip install to attempt to get magenta, but this just led to the same two errors on his machine. I also tried completely wiping python off the computer and it, wiping anaconda off the computer and reinstalling it, and I've tested this with different versions of python.

The strangest part to all this was when I originally installed magenta to my first environment, it actually got so far as to compile the code, until it said tensorflow was out of date, which was what got me on this wild chase to find all the correct versions of everything.

I also have tried just simply ignoring the problems after changing versions or even before and just trying to compile code in IntelliJ, but whenever I try to use the environment, it either won't compile anything at all, or has these two errors.

2020-07-17 00:38:44.182011: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'cudart64_101.dll'; dlerror: cudart64_101.dll not found
2020-07-17 00:38:44.182121: I tensorflow/stream_executor/cuda/cudart_stub.cc:29] Ignore above cudart dlerror if you do not have a GPU set up on your machine.

I figured that maybe I am missing something entirely about this whole situation or hopefully it is something very simple that just seems like a bigger problem than it is.

Any help at all will be appreciated and if there are any logs needed I will provide them. I just want to get magenta working for a project that I was planning on attempting and it seems that just installing it will end up being the most time consuming part of the whole project.

from Magenta installation using "pip install magenta" having errors even on fresh environments