magmadnn
magmadnn copied to clipboard
MagmaDNN
MagmaDNN: a simple deep learning framework in c++
Hi, greate work！ I've successfully compiled the across win/linux platform. And I tried to do some modification to the current version that related to the GPU memory management issue. How...
Added Spack build package for MagmaDNN. [TODO] +OneAPI variant fails at CMake step [TODO] +cuda, +̶m̶p̶i̶, +magma, +̶o̶p̶e̶n̶m̶p̶ are all untested (I don't have access to GPUs right now) To...
**Is your feature request related to a problem? Please describe.** Currently MagmaDNN convolutions are only supported on the GPU. This disallows any testing of networks on CPU only installs. **Describe...
Add Pooling2d support on CPU. Should include max and average pooling. Initially it doesn't need to be optimized. That can be a separate issue.
Finish adding in CPU conv_grad calculation in math files.
**Is your feature request related to a problem? Please describe.** Optimizers just run a single step when the minimize function is called. This "works", but is counter-intuitive. The minimize function...
**Is your feature request related to a problem? Please describe.** Models are not distributed in parallel by default (i.e. when they exceed device memory). Thus the user must first encounter...
**Describe the bug** Passing stack allocated tensor objects to some MagmaDNN functions causes stack smashing errors. The bug is somewhat unpredictable. It does not happen every run. **To Reproduce** ```c++...
Some sort of pipeline abstraction for defining the data load/train/inference pipeline. Perhaps some capability to load and export pipelines. This is low priority, but would make using the framework easier.