search

Elsaam2y / DINet_optimized

An optimized pipeline for DINet reducing inference latency for up to 60% 🚀. Kudos for the authors of the original repo for this amazing work.
93 stars 15 forks source link

Can Dinet fine tune the model? How is it operated? #19

Closed einsqing closed 4 months ago

einsqing commented 4 months ago

Can Dinet fine tune the model? How is it operated?

Elsaam2y commented 4 months ago

You can fine-tune using their own data. You mainly need to continue from the last saved checkpoints and after preprocessing the new data. However, this might not always lead to get good results. I would instead suggest adding the data to HDTF and proceed with the training.