error code 500

#54
by Araeynn - opened

Also, why does the space error with 500 frequently?

Core ML Projects org

Fair question. This space uses a Mac server as a backend, which is an exception in our infrastructure. Everything was set up manually, and monitoring is not as comprehensive as automated as in the rest of the cases. The actual causes of server error are diverse: out of memory situations that kill the process, server reboots (updates or any other reason), etc. We've improved error handling since the initial versions, but we should create a better monitoring and alerting system. It's in my to-do list, I hope to get to it soon :)

May I ask what's your use case, since you appear to be a frequent user of the space? Do you frequently convert models? What types?

My use case specifically is porting multiple models over to coreml and then uploading them to a database.
I have made a MacOs/iOS llm and image gen app for a school project, and I have a model selector which allows users to select which model they would like to use.

I am also converting models frequently to work on my Mac, and I really enjoy the 'Transformers-to-Coreml' space as well as it gives quite good results on the models without having the hassle of tuning with Coremltools. I wonder if it would be possible to release the latest source code so that I can run it on my own computer rather than getting these 505 errors?

I have the code from the commit the 31st of March but reckon that there has been further development since then.

Thanks for great work Pedro!

Sign up or log in to comment