search

PaulKMandal / FedCL

Federated Deep Learning in C and OpenCL
Other
1 stars 0 forks source link

Create Matrix Multiplication Function #4

Closed SilverArgente closed 11 months ago

SilverArgente commented 11 months ago

Create a matrix multiplication function to be used for forward pass of dense inputs (size of matrix is variable)

I will take a look at this issue, will keep posted

PaulKMandal commented 11 months ago

I ended up writing the code for this and rewrote the forward() implementation for Dense. Because forward now takes parameters, I have to rewrite some other stuff before getting implemented. I also need to figure out whether some of the logic of the DAG needs changing.

I'm going to leave this issue open until I push.

PaulKMandal commented 11 months ago

Added in this commit.

We are going to need to think about how we handle the forward pass of the network. We'll also need to add batching. However, those are irrelevant to this issue so I'm closing this for now.