[Bug]: Wrong gradients in the CUDA implementation of Layer Norm about dlib HOT 10 CLOSED

Yes, I will try to figure out. I will check what test the test_layer is doing to see why is not passing... The CPU version passes, but not the GPU version. But then when the test for equality in CPU and GPU passes, too, which is really odd.
Maybe rewriting the whole thing from scratch is easier than trying to figure out the typo, haha.

from dlib.

Yeah I noticed this a bit ago but didn't figure it out when I looked. Need to look again though and figure it out. Although maybe you will figure it out first :) Must just be some subtle typo somewhere.

from dlib.

Yeah hard to say, bugs can be anywhere and you never know until after the fact 🤷 :D

from dlib.

It's really weird. If I change this line:

resizable_tensor x(4, 2, 2, 4);

Then, this one fails: the values are entirely different. But with the previous size, the values are the same.

from dlib.

I will tackle this again at some point. Sometimes, when things like this happen and I have absolutely no idea why, I question myself a lot...

from dlib.

Na everyone thinks that from time to time. Don't sweat it :)

from dlib.

I could be naive here, is there a reason why Layernorm isn't using CUDNN ?
Will cudnnNormalizationForwardInference, cudnnNormalizationForwardTraining and cudnnNormalizationBackward work ?
It looks like those functions can be used for batchnorm, layernorm and groupnorm.

from dlib.

Oh, I wasn't aware it was possible to do Layer Normalization with cuDNN.
This link says that: https://docs.nvidia.com/deeplearning/cudnn/api/cudnn-ops-library.html#cudnnnormalizationforwardtraining is deprecated in v9.
Could you please show me how to use it for LayerNorm? I can't see any mention in the API.

It seems weird that this layer, now used in most Transformer-Based networks, has no cuDNN implementation…

from dlib.

Somewhere in the docs I read it could be used for multiple types of normalization...
I agree, it's hard to believe cudnn doesn't have first class support for it. Maybe libraries like Pytorch, GGML etc have moved away from cudnn and just use vanilla CUDA or https://github.com/NVIDIA/cccl to write their kernels from scratch. Don't know.
Also a lot of Transformers use different normalization functions now, like RMSNorm for example.

from dlib.

Ah, maybe you're referring to this? This is for FC or Conv mode in Batch Norm, which dlib already uses.
https://docs.nvidia.com/deeplearning/cudnn/api/cudnn-ops-library.html#cudnnnormmode-t

from dlib.