Owen L

Unable to recreate, `pip install tensorflow-quantum` just worked for me on a brand new environment. Results in `Successfully installed cachetools-4.2.4 cirq-core-0.13.1 cirq-google-0.13.1 duet-0.2.7 google-api-core-1.21.0 google-auth-1.18.0 googleapis-common-protos-1.52.0 networkx-2.8.8 protobuf-3.17.3 sympy-1.8 tensorflow-quantum-0.7.2`

Colab silently upgraded default python to 3.10.11 I think. It has broken a number of people's codes and I have heard a number of people encountering similar issues for a...

Do you want to run on a simulator or hardware? Your question says you want to use simulators to train on hardware (which doesn't make sense). I assume you want...

How does replacing backend=cirq.Simulator() change the training of tfq.layers.ControlledPQC? It makes it run on cirq instead of qsim. It also runs a lot slower, but there is more customization you...

Yes, qsim is just the default

I don't think GPU ended up being added at all (see the open PR: https://github.com/tensorflow/quantum/pull/687). That being said, if you want to accelerate your code across multiple nodes, you can...

How is this meaningfully different from https://www.tensorflow.org/quantum/tutorials/mnist? Also there is no explanation as to what 6G even is.

My first question meant, what TFQ features/functionality is different that the MNIST example? It is a downsampled image that is encoded via X gates that is then compared to a...

It is important for tutorials to be highly explanatory (both of TFQ and of QML), so all the information that you are telling me should probably be in the tutorial...

As long as you sufficiently attribute credit (i.e. cite properly) that is usually fine. When it comes to MNIST images, those aren't owned by Tensorflow. I don't know who owns...