Runtime environment FAQ (frequently asked questions)

How to save the dependencies I have already installed without having to reinstall them every time I restart?

There are two ways to save your dependencies::

1. By default conda Used in the environment pip install --user xxx Command installation dependency. The dependencies installed in this way will be saved in the user directory. Will not be lost due to restart.
2. Create a custom conda environment. And place the environment in the user directory (/hyperai/home), More information can be found at use Conda Perform dependency management.

But for the passage apt The installed dependencies cannot be saved. Reinstall when startup is required.

Can I use a custom runtime (image)?

Currently, platform users are unable to directly use custom images. But custom frameworks can be used in the corresponding images (framework)

We have already added it TensorFlow. PyTorch. MXNet. Darknet Common versions of preset images are available for direct use. More information about mirrors can be accessed Runtime environment (image)overview Learn more information.

ours Private deployment Services can provide customized mirror services for enterprises.

Is it provided Caffe2 The support?

according to Caffe2 of Official Announcement, Caffe2 Already merged into PyTorch. So we will no longer provide Caffe2 Framework support, at present HyperAI The platform already supports PyTorch frame