This Channel should get more subscriber.. what a learning curve after watching your videos.. I could not stop myself to finish all the lectures in the playlist....Content is so interesting and To the point... Hats off to you Nitish Sir..
aap ne bola ki image size pr depend nahi krta . . . . but jab sare conv layers khatam ho jayenge tab toh flatten hi krna padenga .but ig aap ye bol rahe the ki agar acche se apan ne zyda conv layers rakhe aur pooling bhi kiya , toh image ka size itna zyda matter nahi krta as comapred to ann . BTW a question . . . kya apan some how dense layer ki input ( conv ka o/p ) ko variable rak sakte?
I hats off your conceptual clarity... Each and every thing you taught with very clarity ... that why it helps to understand the concepts very clearly ..............
I couldn't get the first answer correct but got the second answer correct since the parameters change only when there is a change in the number of filters.
Sir pls make a video on k fold cross validation Sir maine abhut saara video dekha par usme acche se explain nahi Kiya gaya hai Aap ek video banao aur statistics par bhi video banao Apne playlist me dusre ka video daala hai Sir it's very important for me I am watching only your video , getting understand with your video only Humble request from honest student to best teacher
Each layer ka 9 weights so 3 layers ka 27 weights and there's total of 50 such filters so learnable parameters must be 27*50 = 1350 and 50 must be biases
ANN vs CNN: similarities and difference learnable parameters doesn't depend on image size in cnn but depend only on no of filters and its size => computational cost reduce
But in CNN there is FC layer, if the image size increase then numbers of trainable parameters also increased in CNN. Can anyone please help me with it?
This Channel should get more subscriber.. what a learning curve after watching your videos.. I could not stop myself to finish all the lectures in the playlist....Content is so interesting and To the point... Hats off to you Nitish Sir..
Happy New Year Nitish Sir 🎊🎆
number of values in filter = 3*3*3*50 = 1350
number of bias values = 50
total number of trainable parameters are = 1400
aap ne bola ki image size pr depend nahi krta . . . . but jab sare conv layers khatam ho jayenge tab toh flatten hi krna padenga .but ig aap ye bol rahe the ki agar acche se apan ne zyda conv layers rakhe aur pooling bhi kiya , toh image ka size itna zyda matter nahi krta as comapred to ann . BTW a question . . . kya apan some how dense layer ki input ( conv ka o/p ) ko variable rak sakte?
You are explaining the things in a very easy manner
Please never stop doing it
Physics Wallah of Deep learning Love you sir g
Best playlist on deep learning...
number of values in filter = 3*3*3*50 = 1350
number of bias values = 50
total number of trainable parameters are = 1400
Amazing vdo.... Nitish sir... Love you 😍😍
the best lecture of my lilfe
U r doing greate ☺️☺️☺️☺️ luv u sir 🥺
Trainable parameter is 3*3*3 *50 +50 = 1400 weights and bias
I hats off your conceptual clarity... Each and every thing you taught with very clarity ... that why it helps to understand the concepts very clearly ..............
Thank you so much for the amazing content. one small request sir, It's more helpful if video upload frequency is increased.
thank you Sir
the concept became clear.
thanku sir from my heart!
one learneable parameter the filter values.
Sir jaldi se DL ke video bana digeye.. Wait kena muskil ho gaya hai... Aap k alawa aur kisi ka video nhi dekh sakta ho
I never said Thank You... 🤝! (The explanation is not just simple but somehow it is also pretty easy to memorize!)
I couldn't get the first answer correct but got the second answer correct since the parameters change only when there is a change in the number of filters.
Beautiful content and extremely well presented.
you are awesome teacherrrr
best playlist
Meri ex ka naam bhi Ann tha Bhai chodke chali gyi......yaha naam daalke aapne yaadein taza kardi......Barish bhi vapas aagayi par Ann vapas nhi aayi 🥲
U r great sir...
Sir please end this series as soon as possible
3*3*3*50 + 50
Sir Thanks a lot for this video .please continue your python 100 days series.Also any update on your live courses. Kindly let us know 🙏
Thank you sir
I have one doubts please reply :- fully connected layers will also have weights apart from weights on filters ???
can you explain 1x1 layer as fully connected ?
sir, when is the mentorship program going to announce?
nice explain sir
Sir pls make a video on k fold cross validation
Sir maine abhut saara video dekha par usme acche se explain nahi Kiya gaya hai
Aap ek video banao aur statistics par bhi video banao
Apne playlist me dusre ka video daala hai
Sir it's very important for me I am watching only your video , getting understand with your video only
Humble request from honest student to best teacher
Thanks
Sir, the answer to your question would be 1350 + 50 (bias) = 1400 trainable parameters
Each layer ka 9 weights so 3 layers ka 27 weights and there's total of 50 such filters so learnable parameters must be 27*50 = 1350 and 50 must be biases
lectures are really good
excellent explanation
learnable parameters = 3*3*3*50+50 = 1400
there will be 1400 trainable parameters 1350 parameters for the 50 filters and the 50 biases of each filter
at time 12:00:-It will have 1400 trainable parameters
The 2nd question answer is the same as 1400
Enjoy your teaching style, all things to the point.
Hello sir, I want to know that how to implement WEKA regression model into mobile app.
ANN vs CNN: similarities and difference
learnable parameters doesn't depend on image size in cnn but depend only on no of filters and its size => computational cost reduce
Sir the answer is 1400 learnable pararmeter because...3*3*3*50 = 1350 filter values to be learned and 50 bias for each filter == 1400
Rgb image so
We 9x3 (9*3) and 50 biass (kernel size ) and than actual ans is »
9*3*50= 1350 .
total trainable parameter will be: 1400
Beautiful
Sir , I completed your ML And DL all videos and waiting for more video
your LinkedIn profile please!
But in CNN there is FC layer, if the image size increase then numbers of trainable parameters also increased in CNN. Can anyone please help me with it?
I have the same question 😮
Amazing
remarkable
Awesome 👍
same
Make a video on unet model
SIr, Please start object detection algorithms "SSD, R-CNN, YOLO..."
Oh.... It's 3rd time.... 1st comment.... 😆😆😆
lusm sir
maja aa gaya sir
finished watching
1350 weights + 50 bias
28*50
1400 are learnable parameters
1400 learnable
best
1350+50(bias)
Its 1400 Lerneble Parameeter
1400
1350
3*3*3 * 50 + 50
Request for RNN/ LSTM-/GAN AUTOCODERS . - Algorithm
1350
1400
1400
1400