Siddharth Venkatesan

This is the expected behavior for PyTorch - by default if a PyTorch Tensor (or in this case a DJL NDArray backed by the PyTorch engine) enables`setRequireGradient`, the gradients are...

I'm glad you found a way to get around the issue - that solution certainly works @ajrnz. I'll take a look at the D2L book and see where a good...

@KexinFeng we could, but I don't think we should. In case someone is using DJL + PyTorch for distributed training, they may choose to accumulate gradients over some microbatches before...

This functionality is now included in DJL https://github.com/deepjavalibrary/djl/blob/master/engines/pytorch/pytorch-engine/src/main/java/ai/djl/pytorch/jni/IValue.java#L313.

@mehmetcalikus we have recently enhanced the DJL Huggingface Tokenzier implementation with lots of new features including padding and truncation, as well as an updated model zoo. The updated Tokenizer should...

We now have a Yolo imperative model [YoloV3](https://github.com/deepjavalibrary/djl/blob/master/model-zoo/src/main/java/ai/djl/basicmodelzoo/cv/object_detection/yolo/YOLOV3.java) in our model zoo.

Closing due to inactivity. Please reopen if you still have questions.

Closing due to inactivity. Please reopen if you still require help.

Hi @mehmetcalikus - I'm currently working on exposing decode and batch_decode functions in java. Will keep this thread updated.

@mehmetcalikus The tokenizer.decode(...) method has been added to DJL. You should be able to pull it from the latest SNAPSHOT. Let me know if you have any issues using it.