Mike McCann
Mike McCann
Many LinOps have tests for scalar multiplication, i.e., that a * (H @ x) = (a * H) @ x. Right now, this involves code duplication, e.g., https://github.com/lanl/scico/blob/1a66887a03ad4cca5dfa58dfa86386f1536926ba/scico/test/linop/test_circconv.py#L74-L87 versus https://github.com/lanl/scico/blob/1a66887a03ad4cca5dfa58dfa86386f1536926ba/scico/test/linop/test_diag.py#L146-L158...
Proposed `input_shape` and `output_shape` policy: * All instantiated `Operator`s must have `input_shape` and `output_shape` properties * For an `Operator` `H`, `H(x)` should throw an error if `x` is not `input_shape`...
Arithmetic operations (+, -, *, ...) between a SCICO BlockArray and a NumPy ndarray are defined in the [BlockArray docs](https://scico.readthedocs.io/en/latest/_autosummary/scico.blockarray.html#module-scico.blockarray) to work in two cases: (1) when the ndarray is...
We should decide what `input_dtype` and `output_dtype` mean, then document it. Proposed `input_dtype` and `output_dtype` policy: - All instantiated `Operator`s must have `input_dtype` and `output_dtype` properties - Possible data types...
In _real scientific imaging code_ one of our team members has wanted to sum each block of a block array along 2 of their 3 dimensions. Can we extend the...
JAX works in single precision by default, and won't even let you create double precision arrays unless an environment variable is set (`JAX_ENABLE_X64=True`) or a special command is run when...
See example results at https://scico--529.org.readthedocs.build/en/529/examples/ct_projector_comparison_3d.html Still not clear why the back projections don't look like they match. It seems like calling `.transpose(2, 1, 0)` on the ASTRA back projection gets...
resolves #542 hypothesized to resolve a warning about dtypes from the tests
### Description I get a segfault when running the following code snippet on my GPU machine. ```python import numpy as np import jax import jax.numpy as jnp def f_host(x): #...