Hi Amey. You are right. Just for the sake of introducing "t" as an expansion factor of the input, it is better to consider that the dimensions of the pointwise kenel is 1×1×d×dt, where the quantity "dt" is the number of kernels and the "t" in "dt" is again the expansion factor. I'll review this video again and try to warn the viewers about it. Thanks a lot and sorry if I caused you any confusion 🙏 .
I go over the entire web and didn't find an explanation for MobileNetV2 except your videos Your explanation is great, detailed, and easy to understand thank you very much
it is better to consider that the dimensions of the pointwise kenel is 1×1×d×dt, where the quantity "dt" is the number of kernels and the "t" in "dt" is again the expansion factor
Shouldn't the output of pointwise conv h x w x t? why do you say it is h x w x dt?
Hi Amey. You are right. Just for the sake of introducing "t" as an expansion factor of the input, it is better to consider that the dimensions of the pointwise kenel is 1×1×d×dt, where the quantity "dt" is the number of kernels and the "t" in "dt" is again the expansion factor. I'll review this video again and try to warn the viewers about it. Thanks a lot and sorry if I caused you any confusion 🙏 .
I go over the entire web and didn't find an explanation for MobileNetV2 except your videos
Your explanation is great, detailed, and easy to understand
thank you very much
I appreciate your feedback. Thank you
it is better to consider that the dimensions of the pointwise kenel is 1×1×d×dt, where the quantity "dt" is the number of kernels and the "t" in "dt" is again the expansion factor
please read my reponse to the pinned comment. I believe I had answered there the same question. Let me know if it's not the answer you want.