autograd

🚀 Feature

Teach torch.Tensor.scatter_ to handle index.size(d) > src.size(d).

Motivation

Currently, torch.Tensor.scatter_ requires index.size(d) <= src.size(d) for all dimensions d, unless src is float-valued. This constraint seems artificial.

import math
import torch

device = 'cuda'  # fail

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

We would like to define a new operation called SISWAP, that is the square-root of the qml.ISWAP operation. See further details here.

Implementation

List of tasks for completion:

A. Adding the operation

1. Adding the SISWAP new class for the operation to pennylane/ops/qubit/non_parametric_ops.py. Complete its

Following the merge of the adjoint code, our coverage fell dramatically. We should put in some effort to get it back to 100%! :-)

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