Hey. I am considering using GANs for my data augmentation to tacke imbalanced classed in the task of Facial Emotion Recognition. I am planning to use Google collab to train my GAN model with datasets contain specific emotion. What gan model would do best in creating most realistic results? CycleGAN?

Hi Aladdin, I have a Query that how you are able to calc the IOU scores from just length and breadth of the bounding boxes, can you please explain me that?

Thx.

Thanks for sharing 😀👍 greetings from Colombia.

At 54:02, why do we need to pass the truncated trg[:, :-1] ? I thought we already applied the mask to prevent the model from looking into the future?

I think the confusion people may have is that the bounding boxers can be different sizes and you might have an excessively big predicted bounding box into a small actual.

Walker Christopher Miller Sarah Gonzalez William

I am new to deep learning should i listen to this videos or go to deep learnng playlist and what is the next step i should folow?

Davis Edward Robinson Michelle Johnson Larry

which color theme isn it?

Lewis Michael Lee Sarah Thomas Betty

Thanks for sharing , that's really helpful!

Samir Harbor

I recoded this in C++. It was interesting

there is a simpler way to reduce lr you can import ReduceLROnPlateau

Also love your voice btw, feels calming 👍

Thanks so much!🎉

After 17 submission, I'll watch your video sir!

I love the way your ide looks, what are you using/ what settings?

10:03 I think the sigma here, which is the output of "self.hid_2sigma", should be interpreted as "log of variance". Why? Because the Linear layer can somehow output negative values while Variance (of a distribution) can't be negative. By interpreting as "log of variance", we can get the variance inside by using exp(). As a result, we have two needed code changes: doing the exp() against this "log of variance" when calculating z as well as calculating the Loss.

Conor Walks

how the discriminator, at the start , knows to distinguish for instance a dollar bill?

Hi! thanks for your explanations. It could be in tensorflow also? Thanks!

Walker Ronald Perez Michelle Lee Carol

i want to learn tensor flow with c++ make tutoreials on that too

I'm trying with tensorflow 2.17.0 and the line inputs = keras.Input(shape=(28*28)) produces the error "cannot convert '784' to a shape. solution: apparently shape has to be explicitly a tuple, so shape(28*28, ) (with comma) does it.

You are doing such an incredible join man

thanks for this! your guide is clear and makes sense

15:26 i don't think it is good to shuffle the test data, if you want to compare models bassed on test you need should not shuffle the test.

pls work on the scratch implementation

what is this ide ?

why would you ever undervolt lol thats like getting a lambo and putting it in eco mode

can you please share the code

Wow maan, it was really an impressive experience

If you get the error 'ValueError: Argument(s) not recognized: {'lr': 0.001}', the issue is because 'lr' has been deprecated in favor of 'learning_rate'. Thus the code should now be: optimizer=keras.optimizers.Adam(learning_rate=0.001),

Very High Temperature can damage your GPU.

thanks

Could you please do a video on the compute hardware. Like the nvidia graphics cards or AMD, where you estimate model parameter training capability.(Provided that the model is a straight forward model like a transformer) or something like that.

your voice is very cool and soothing

For some reason some of the graphs are totally fine and some show up very limited step.

谢谢你的视频

hope you feel great

on the apple watch you can enable track af fibirilaltions which gives automatic hrv readings i think hourly also during the day

Omg its so funny I was literally about watching a review about the whoop and now you dropped one

Hi. First of all really appreciate the great guide you've uploaded for undervolting the GPU. However, I have a question running in my mind. Hope you aren't busy to reply back, but why did you lower the core clock by - 250 before proceeding to change the voltage curve? Is there some specific reason for this?

14:05 is this even correct though? Shouldn't you project Q, K and V from the `embed_size` to `embed_size` and split the resulting tensors? Your implementation seems to treat parts of the embedding dimension separately (instead of jointly attending to the entire embedding space) and using the same weights (instead of having separate weights for each head).

Doesn't line 68 loss_disc.backwards() cause the generator to train at the same time unless you freeze the weights? I was under the impression the generator is not supposed to train at the same time as the discriminator because then you are making the gen worse to improve the disc.

Does this support writing to remote path directly?

at 9:30 he says the discriminator wants to maximize the loss? shouldn't it be wanting to minimize it? can someone help me understand? TIA

6:58, 'that looks AWFUL' was best 🤣