Hey there Hopsworks team, incredible works being done here demistifying things like training vector embeddings models and two tower recommendation systems.
Although the youtube talks from Jim are very nice to watch on building the tiktok real time systems, most of the code he uses on the presentations have now changed and are a bit confusing to follow. I would love to follow a written step by step tutorial on running the tiktok example at https://github.com/logicalclocks/hopsworks-tutorials/tree/branch-4.2/real-time-ai-systems/tiktok_recsys
The file system is kinda complex, you have java, jupyter, and also those folders goes deep down into many others more - features, streaming, utils, etc. Then theres the enumerated jupyter files 1a, 1b etc.
Anyway, got a bit lost when running the tutorial for using hopsworks for doing this real time recommendation systems on where should I really start, to understand what's really going on and work on a proper implementation of it.
Would be awesome if you guys could give some directions or update the readme in that section of the repo.
Thanks!
Hey there Hopsworks team, incredible works being done here demistifying things like training vector embeddings models and two tower recommendation systems.
Although the youtube talks from Jim are very nice to watch on building the tiktok real time systems, most of the code he uses on the presentations have now changed and are a bit confusing to follow. I would love to follow a written step by step tutorial on running the tiktok example at https://github.com/logicalclocks/hopsworks-tutorials/tree/branch-4.2/real-time-ai-systems/tiktok_recsys
The file system is kinda complex, you have java, jupyter, and also those folders goes deep down into many others more - features, streaming, utils, etc. Then theres the enumerated jupyter files 1a, 1b etc.
Anyway, got a bit lost when running the tutorial for using hopsworks for doing this real time recommendation systems on where should I really start, to understand what's really going on and work on a proper implementation of it.
Would be awesome if you guys could give some directions or update the readme in that section of the repo.
Thanks!