Refactoring kernels #206
Refactoring kernels #206
Conversation
Codecov Report
📣 This organization is not using Codecov’s GitHub App Integration. We recommend you install it so Codecov can continue to function properly for your repositories. Learn more @@ Coverage Diff @@
## refactor_kernels #206 +/- ##
====================================================
+ Coverage 28.29% 28.63% +0.33%
====================================================
Files 70 70
Lines 3262 3332 +70
====================================================
+ Hits 923 954 +31
- Misses 2339 2378 +39
Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here. |
gpjax/kernels/stationary/matern52.py
Outdated
| self, params: Dict, x: Float[Array, "1 D"], y: Float[Array, "1 D"] | ||
| self, x: Float[Array, "1 D"], y: Float[Array, "1 D"] |
There was a problem hiding this comment.
We need to replace all kernel inputs of Float[Array, "1 D"] with Float[Array, "D"] for the x's and y's!
gpjax/kernels/stationary/white.py
Outdated
| class White(AbstractKernel): | ||
| def __init__( | ||
| self, | ||
| compute_engine: AbstractKernelComputation = ConstantDiagonalKernelComputation, | ||
| active_dims: Optional[List[int]] = None, | ||
| name: Optional[str] = "White Noise Kernel", | ||
| ) -> None: | ||
| super().__init__(compute_engine, active_dims, spectral_density=None, name=name) | ||
| self._stationary = True | ||
|
|
||
| variance: Float[Array, "1"] = param_field(jnp.array([1.0]), bijector=Softplus) |
There was a problem hiding this comment.
For the White kernel, the default computation should be ConstantDiagonalKernelComputation , so that
from simple_pytree import static_field
@dataclass
class White(AbstractKernel):
variance: Float[Array, "1"] = param_field(jnp.array([1.0]), bijector=Softplus)
compute_engine: AbstractKernelComputation = static_field(ConstantDiagonalKernelComputation) # <- set the default| def test_gram(self, dim: int, n: int) -> None: | ||
| kernel: AbstractKernel = self.kernel() | ||
| kernel.gram | ||
| x = jnp.linspace(0.0, 1.0, n * dim).reshape(n, dim) | ||
| Kxx = kernel.gram(x) | ||
| assert isinstance(Kxx, LinearOperator) | ||
| assert Kxx.shape == (n, n) | ||
| assert jnp.all(jnp.linalg.eigvalsh(Kxx.to_dense()) > 0.0) |
There was a problem hiding this comment.
When you test positive definiteness, I would recommend adding a small jitter, to the diagonal of the gram matrix Kxx, I found that one of the tests failed for me locally on my machine.
I would also give a quick comment to say what the line:
assert jnp.all(jnp.linalg.eigvalsh(Kxx.to_dense()) > 0.0)is doing to make it clear for others!
Just contributing to #199.
Pull request type
Please check the type of change your PR introduces: