pip | a web server for gentlemen | HTTP library

Because HappyTransformer does not support multi class probabilities I suggest to use another library. The library flair provides even more functionality and can give you your desired multi class probabilities, with something like this:

I managed to do it with source distributions and overriding sdist and egg_info commands to make setuptools bundle local dependencies together with package and to make pip search dependencies in that bundle when installing the built package later. But later I figured out it makes system vulnerable to dependency confusion attacks because local packages installed from that bundle are visible with pip freeze, if for some reason the dependency location, like local-package @ file:///home/user/packages/local-package.tar.gz is stripped to just local-package pip will search it on pypi, which allows dependency confusion to happen.

The best solution for this problem is to vendor all local dependencies where their source code is copied to the package, pip itself vendors its dependencies using vendoring.

I resorted to using the community module for elastic ip. So community.aws.ec2_eip instead of ec2_eip.

I eventually scanned through this one below, that although it's for Windows it actually worked on Ubuntu linux too!!

Pip ReadTimeoutError in Windows 10

and the way to fix it is then simply disable ipv6 with the following commands, and done!

I would suggest you to use conda (Ananconda/Miniconda) to create a separate environment and install tensorflow-gpu, cudnn and cudatoolkit. Miniconda has a much smaller footprint than Anaconda. I would suggest you to install Miniconda if you do not have conda already.

Inside your requirements.txt change scipy line with this scipy==1.6.0 and save. Now retry pip installation.

I think when you are using the conda environment. The conda list is going to show all the general packages that shared by the same conda environment. And the reason why 'tensorflow-gpu' is listed in 'pip freeze', but not in 'conda list', is because you used pip install to installed 'tensorflow-gpu'(could be you or the IDE). In this case, 'tensorflow-gpu' is only exists under this python project I believe. Actually, there is an official document about this topic.

Issues may arise when using pip and conda together. When combining conda and pip, it is best to use an isolated conda environment. Only after conda has been used to install as many packages as possible should pip be used to install any remaining software. If modifications are needed to the environment, it is best to create a new environment rather than running conda after pip. When appropriate, conda and pip requirements should be stored in text files.

Use pip only after conda Install as many requirements as possible with conda then use pip.

Pip should be run with --upgrade-strategy only-if-needed (the default).

Do not use pip with the --user argument, avoid all users installs.

And here is the link.

You can use numpy.mean() with axis=0 on the numpy.array() consisting of the closing prices of current date plus 2 days ahead to get the mean, as follows:

1. It's not necessary to use .groupby because there is only 1 record per day.
2. Don't use .reset_index(0, drop=True) because this is dropping the date index. When you drop the index from the calculation, it no longer matches the index of df, so the data is added as NaN.
   • df['Close'].rolling(window=days_back, min_periods=days_back).apply(get_slope, raw=True) creates a pandas.Series. When assigning a pandas.Series to a pandas.DataFrame as a new column, the indices must match.