autograd

Motivation

Currently lots of C++ based unit tests are executed directly from test.sh/win-test.sh for example:
https://github.com/pytorch/pytorch/blob/0bd8d0951dcb4063c0f7552a7404bd7f0e7b6e6f/.jenkins/pytorch/test.sh#L317

Which have following drawbacks:

• It excluded those test runtime from auto-sharding/auto-categorization
• Make them subject of running on only particular platform (

Description

Implement a YOLO model and add it to the DJL model zoo

References

• https://arxiv.org/abs/1506.02640

Issue to track tutorial requests:

• Deep Learning with PyTorch: A 60 Minute Blitz - #69
• Sentence Classification - #79

Feature details

Update handling of sort in qml.math such that the return values are the same across all interfaces.

The current behaviour is as follows. For numpy, jax, and tf,

>>> a = np.array([1, 3, 4, 2])
>>> qml.math.sort(a)
tensor([1, 2, 3, 4], requires_grad=True)

But for torch, the argument indices are also returned:

>>> a = torch.tensor

Add a function and module that permits spike-time decoding, as suggested by @schmitts https://twitter.com/sbstnschmtthd/status/1432343373072019461

The init module has been deprecated, and the recommend approach for generating initial weights is to use the Template.shape method:

>>> from pennylane.templates import StronglyEntanglingLayers
>>> qml.init.strong_ent_layers_normal(n_layers=3, n_wires=2) # deprecated
>>> np.random.random(StronglyEntanglingLayers.shape(n_layers=3, n_wires=2))  # new approach

We should upd

Okay, so this might not exactly be a "good first issue" - it is a little more advanced, but is still very much accessible to newcomers.

Similar to the mygrad.nnet.max_pool function, I would like there to be a mean-pooling layer. That is, a convolution-style windows is strided over the input, an