djl
djl copied to clipboard
An Engine-Agnostic Deep Learning Framework in Java
## Description ND4J, the accelerated linear algebra backend that powers Eclipse DeepLearning4J, should have all the necessary features to become a proper backend for DJL. ### Will this change the...
## Description At the moment every web-client/browser needs access through firewalls/proxies to download the model using central from the original URL as well as to central. Instead of that it...
Introducing an optional output `NDList` parameter on the `Predictor::predict` interface. When the output parameter is provided the underlying engine will copy the inference result into the corresponding `NDArray` instead of...
https://github.com/awslabs/djl/blob/master/model-zoo/src/main/java/ai/djl/basicmodelzoo/cv/classification/ResNetV1.java ``` public static Block residualUnit( int numFilters, final Shape stride, final boolean dimMatch, boolean bottleneck, float batchNormMomentum) { SequentialBlock resUnit = new SequentialBlock(); if (bottleneck) { resUnit.add( Conv2d.builder() .setKernelShape(new...
## Description There may be some bug in the json code which causes errors when the server populates the model view fronted. This causes some model attributes to not show...
## Description TF engine can only be used for inference, when can be used for trainning? ### Will this change the current api? How? Yes. Wrap TF java api 2.x....
## Is there a way to use torchscript model with custom ops? I'm trying to run pytorch geometric model with DJL and have a problem with custom ops from pytorch...
## Description Developers will be able to create an auto-generated template through the cli or frontend interface, or manually enter the model information through the ModelZoo webUI, so it can...
## Description An addendum to enhancement #634. Removing Models will be the another necessary feature along with Model Uploading. This will also require a button on the ModelView page.