Vision Transformer explained in detail | ViTs

[@soravsingla8782]

Your videos are always unique & highly knowledgeable. Thank you

[@CodeWithAarohi]

Thank you!

[@layamahmoudi6002]

Thank you for the amazing video, it's absolutely perfect!

[@CodeWithAarohi]

I'm glad you found it helpful!

[@TruthOnly_jayshreeRam]

awesome, very nicely explained. Thanks Ma'am.

[@CodeWithAarohi]

Most welcome 😊

[@TruthOnly_jayshreeRam]

@@CodeWithAarohi Ma'am, Can you please make a video on "memory-augmented zero-shot image captioning"

[@vcarvewood4545]

You are excellent teacher. I'm in love in your voice since YOLOv8 tutorials. Attention to Aarohi is all we need.

[@CodeWithAarohi]

Thank you for the compliment! I'm really glad the tutorials and my voice have made learning enjoyable for you.

[@revanb2781]

Very impressive video, Thank you.

[@CodeWithAarohi]

You're welcome!

[@arnavthakur5409]

Excellent content ma'am

[@CodeWithAarohi]

Glad you found it helpful!

[@Sunil-ez1hx]

Such an informative video🙏🙏

[@CodeWithAarohi]

Thanks, glad you found it helpful!

[@eranfeit]

Thank you for a great explanation

[@CodeWithAarohi]

You are welcome :)

[@AsthaPatidar-w1t]

Please make a video for Convolution to Vision Transformer in detail. And thanks for this video.

[@CodeWithAarohi]

Noted!

[@pifordtechnologiespvtltd5698]

Extremely Appreciated 👏👏👏

[@CodeWithAarohi]

Thank you so much 😀

[@bharatto2220]

Thankyou for explaining the videos very elaborately and clearly. But at some places it was too basic like RGB, would appreciate a timeline so that I can skip to the required part

[@CodeWithAarohi]

Thankyou and I will add timeline

[@aneerimmco]

informative, thank you ma'am.

[@CodeWithAarohi]

Most welcome 😊

[@munimahmed9374]

Can you please explain DEiT model, this Vit explanasion is the best video on Vit I found on the internet. thanks a lot

[@madhavanu6980]

Transformers for remote sensing classification paper ma'am.....plz explain it ma'am...bcz you do it great and in easily understandable manner

[@Ishaheennabi]

it's great video thanks for it

[@CodeWithAarohi]

Glad you liked it!

[@satvik4225]

43:20 You said that we do element wise addition of Patch representation and position embedding which means their dimension is same. The patch representation is of length 768x1 and you also said the length of the position embedding vector is 512. How will you do the element wise addition. did you mean linear projected vector of eatch patch which has dimension of 512? I learnt alot of stuff, thanks

[@CodeWithAarohi]

@@satvik4225 Every patch is of size 512x1 after linear embedding layer. And then we are adding position encoding with patches.

[@satvik4225]

Okay, thanks. It was a bit confusing as you said here that patch representation is flatenned patches. Thanks again

[@salmareang7458]

I have some confusion take one input image then how qkv are find ?

[@jynpogger]

God Please Protect My Teacher at all costs

[@CodeWithAarohi]

Thank you so much for your kind words and blessings! 😊🙏

[@ramchandhablani9834]

Mam, your video is very good, I have two questions, If there are 2 hidden layers, then there will be three matrices say W1, W2 and W3 for linear projection. The 2nd question is to train these weights and biases, we neet target vectors corresponding to each input vector. from where we will get those target vectors?

[@CodeWithAarohi]

W1: for projecting image patches into embeddings. W2: for query, key, and value projections in the self-attention layer. W3: for the feed-forward layer after attention. The target vectors come from the labeled dataset, where each input image has a corresponding label (for classification tasks) .

[@soravsingla6574]

Perfect

[@CodeWithAarohi]

Thanks!

[@Mulugeta-c5q]

Thank you for your Good work and can you make a video for ViTPose code too?

[@CodeWithAarohi]

@@Mulugeta-c5q Sure

[@aryarushipathak5039]

Hello ma'am , can we use Vit and CNN to identify emotions from the face ? CNN for feature extraction and mtcnn for emotion labeling

[@nursami7842]

Hello ma'am, can you explain in more detail about encoder transformers such as normalization, multihead attention, softmax, MLP, The video doesn't provide a detailed explanation about that, can you explain that in the next video?

[@nursami7842]

or is there a reference that explains in detail about it?

[@CodeWithAarohi]

Will try to cover in another video.

[@sreenalakhani4985]

pls make a video on video vision transformer also

[@CodeWithAarohi]

Sure!

[@satvik4225]

can you explain diffusion models next

[@CodeWithAarohi]

Noted!

[@madhavanu6980]

Ma'am plz do video on TRS remote sensing transformers...plz ma'am its my humble request....as i completely understand the ViT still I can't understand TRS plz ma'am

[@CollegeOnline]

mam please please please please please please please create video on Gated Vision transformer as i am trying to use it in my research paper, but I am not able to find any literature regarding GVT. mam if you have any links to GVIT then kindly share it please

[@CodeWithAarohi]

Hi, I need time to make this video because I never used this model before and I have read the paper and understand it in order to create a video.

[@fatima-arbab]

Ma'am plzzz make a video Copy move forgery detection in video using machine learning Using yolo model and dataset casia

[@shivamsood4566]

Mam pls also make on stable diffusion,pls mam 👏

[@CodeWithAarohi]

Noted!

[@adityanjsg99]

Arohi ji, possible for you to build a Model which is as good as gpt? Though on limited data and scale..

[@CodeWithAarohi]

@@adityanjsg99 I am not sure 🤔

[@rahulhanot4481]

when will we get more video of this topic

[@CodeWithAarohi]

@@rahulhanot4481 kzbin.info/aero/PLv8Cp2NvcY8AzNCATbDWMr8vqbJBYbxFW&si=-488A3JJRBAVtEdh

[@AB099]

EasyOCR next tutorial text detection

[@CodeWithAarohi]

Noted!

[@khalidamiralam6469]

Hello ma'am, I am still waiting for your video on video generation

[@CodeWithAarohi]

Will upload soon

[@danigunawan6807]

please d your share code explainned thanks ?

[@chinnaiahkotadi4702]

Thank you for your effort mam.but you unable to explain how to position value getting, 2. You not explain while neural network work with help of relu activation functions 3. While making relationships between quer,key and value , what is role of Key

[@CodeWithAarohi]

Noted! Will try to cover in another video.

[@sanathspai3210]

Hi Arohi It was good session. One suggestions could you please make video of all the yolo versions starting from V1 to v11? Many are waiting for it and will be very beneficial

[@CodeWithAarohi]

Hi, Glad my video is helpful! Great suggestion, I will surely make a video on all yolo models.

[@sanathspai3210]

@ please try to tag me once you are done. I’m very much in need of it and the way you explain it

[@CodeWithAarohi]

Sure