qutip-qip backends for qiskit #155
Conversation
claretgrace0801
force-pushed
the
qiskit-backends
branch
from
2184a1a
to
b71f236
Compare
claretgrace0801
force-pushed
the
qiskit-backends
branch
from
0bd7492
to
b708d10
Compare
src/qutip_qip/qiskit/backend.py
Outdated
| if statistics.cbits[0] is not None: | ||
| for i, count in enumerate(statistics.cbits): | ||
| counts[convert_to_hex(count)] = round( | ||
| statistics.probabilities[i], 3 | ||
| ) |
There was a problem hiding this comment.
While displaying the result, the non-rounded probabilities are displayed by default.
There was a problem hiding this comment.
It is better to keep the default as not rounded. Although the rounded one looks cleaner, it unnecessarily makes the result inaccurate. If we do want a better display, we should only round it when displaying, not the original result. But I guess .data() is meant to be raw data and one can define some .probability for nice looking and give it some parameters e.g. truncation.
BTW, why does counts returns a probability... I know that we don't repeat. So we probably don't have a well-defined counts? How does the qiskit circuit simulator handle this?
There was a problem hiding this comment.
Qiskit backends have an option to set the number of shots. I haven't implemented any configurable backend options yet and will be doing that soon. After doing that, the probabilities will be replaced by the actual counts. We can use random sampling from the obtained probabilities for this purpose.
src/qutip_qip/qiskit/converter.py
Outdated
| _ignore_gates = ["id", "barrier"] | ||
|
|
||
|
|
||
| def _make_user_gate(unitary: np.ndarray, instruction: Instruction): |
There was a problem hiding this comment.
Let's make it clear that it is the qiskit Instracution, qutip also has its own Instruction.
|
Seems that codeclimate is undergoing some connection error. I'll just merge. |
This draft PR tracks changes for the GSoC '22 project: making qutip-qip backends for qiskit.