What is better create my own model or use transfer learning ?

[prdai]

What is better create my own model or use transfer learning?
Which will result in a better model?

• Ranuga

[seanbenhur]

Transfer learning is better since the model has pretrained weights which are learned from a large datasets(commonly Imagenet for Computer vision), By using transfer learning, you can save a lot of time in traing the model You can train your model from scratch, if your task is Time Series/Anamoly detection/Tabular Datasets I will suggest you to do Transfer learning, if your goal is to get higher perfomance in your model and the task is related to CV/NLP

…

On Fri, 12 Mar 2021, 17:27 Ranuga-D, ***@***.***> wrote: What is better create my own model or use transfer learning? Which will result in a better model? - Ranuga — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#18>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AKKLL6P64TFNLFKCKOE4KS3TDH6RXANCNFSM4ZCEYSDQ> .

[prdai]

Thanks :)