At 15:50 when I am explaining encoders, I missed speaking about one important point. Note: Unlike previous episodes, both encoders are fine-tuned i.e requires_grad is set to true so the weights are backpropagated. That's how the model learns to tweak the joint embedding space for multi-modal face recognition. The code reflects it, you may pause to look at it.
@seanbenhur
Жыл бұрын
Amazing, I learned a couple of new things!
@prithivida
Жыл бұрын
Thanks Sean
@nandurajendran6986
Жыл бұрын
Hats off for putting great efforts in delivering such high quality content .In today's myriad of data science KZitem channels,only few really sticks with you ,I am happy to say this is one of them .This channel is truly a gem to follow and keep ourselves upto date with the latest trends . A big thank you prithvi.
@prithivida
Жыл бұрын
Thanks for the kind words Nandu, glad to be a signal among noise :). I will work hard to add more value to the community. Thanks again !
@aaearts
Жыл бұрын
Awesome, I learned multiple new things. Truly appreciate the attention to detail to compare 4 models with SoTA
@prithivida
Жыл бұрын
Thanks for the kind words
@rajravip
Жыл бұрын
Thanks a lot for putting so much effort and making the videos of premium quality. Super helpful for experienced professionals as well as beginners.We owe you a lot.
@prithivida
Жыл бұрын
Thank you ! more videos and topics coming in 2023, stay tuned and do share with the community if you think they can benefit.
@FirstNameLastName-fv4eu
Жыл бұрын
This is very good !!! Very beautifully crafted.
@prithivida
Жыл бұрын
Thanks for the kind words.
@vetrivelps6047
Жыл бұрын
Amazing video thanks a lot Prithivi bro 😍
@prithivida
Жыл бұрын
Thanks Vetri :)
@vishalahuja2502
Жыл бұрын
This is super good! Please keep more coming
@prithivida
Жыл бұрын
Thanks Vishal
@credibleparent3883
Жыл бұрын
looking forward to see more.... where are you from in india? a keralite?
Пікірлер: 21