-
This is a feature request.
I am writing a neural net toolkit. Now we can write a convolution function using *im2col* and *GEMM* functions, but *col2im* is also necessary to calculate the back propa…
-
**Chapter name**: Introduction
**List of bugs/typos**
Before any sections
1. “This means that someone **that** thinks of numbers…” should be “This means that someone **who** thinks of numbers…”…
-
Use BLAS Batch computation as a model.
-
Hi,
This is a feature request.
The standard approach of first using im2col and then using gemm for a convolution has the disadvantage that one very easily runs out of memory. There is actually no …
-
Would it be possible to implement Wishart / InverseWishart / LKJ priors?
gpytorch has them [already](https://github.com/cornellius-gp/gpytorch/blob/master/gpytorch/priors/wishart_prior.py), but wh…
-
Hey everyone,
While reviewing this [PR](https://github.com/webmachinelearning/webnn/pull/2) I had an issue but it's a horizontal issue that should be discussed outside of that PR, maybe a good agen…
-
When training with samples of size 256x256 pixels, a batch size over 32 causes a memory error from cuda. We have to find a way to optimize the process, in order to increase the batch size.
```shell…
-
Maybe I'm misunderstanding but if I print both out and labels in the _loss function here: https://github.com/boluoweifenda/WAGE/blob/master/source/NN.py#L96
I get values that are floating point:
`…
-
Hi,
Need some help here.
I do not know how to change the logistic regression example to Multinomial logistic regression.
Example like
Which blood type does a person have, given the results of va…
-
I'm an NLP researcher who does a lot of work with sequence-to-sequence models. Naturally then, my preference is for code that is written in a way that minimizes the time and effort required to impleme…