Michael Haas
Michael Haas
Python 3.5 is EOL. We should remove support for Python 3.5. The deprecation of Python 3.5 support is announced in the 0.8.0 release via #98.
Right now, we only allow binary, which requires additional work compared to just opening a file or passing text.
To be able to re-use a model name, we need to delete both the Model resource **and** the Job ressource. If the Model resource has been deleted, it is not...
As raised by @karthikmswamy on #72.
The current implementation will try to perform inference for all chunks. If an exception occurs, the progress is lost. By yielding the individual chunks, the caller can be in charge...
I have seen this error a few times already: ![Screenshot 2020-07-17 at 09 51 06](https://user-images.githubusercontent.com/260872/87762247-1924d700-c813-11ea-88a2-caad1bf38aad.png) I assume this happens because the timeout is too short while GETing the job. This...
From the current API docs for ModelManagerClient and friends, it is not obvious how to create an instance. This knowledge is hidden in the "internal" documentation for BaseClient.
We only have limited capacity on the live endpoint using for system tests. Instead of using a live endpoint, we can also consider using `pact`.
Fix some typos