jax-ml/jax · error · NotImplementedError

dynamically changing the bandwidth method is not supported

Error message

dynamically changing the bandwidth method is not supported

What it means

gaussian_kde.set_bandwidth is intentionally unimplemented in JAX: bandwidth is fixed at construction because KDE attributes are immutable (set through _setattr) to stay jit-compatible. Calling it always raises NotImplementedError.

Source

Thrown at jax/_src/scipy/stats/kde.py:229

  def logpdf(self, x):
    """Log probability density function"""
    check_arraylike("logpdf", x)
    x = self._reshape_points(x)
    result = _gaussian_kernel_eval(True, self.dataset.T, self.weights[:, None],
                                   x.T, self.inv_cov)
    return result[:, 0]

  def integrate_box(self, low_bounds, high_bounds, maxpts=None):
    """This method is not implemented in the JAX interface."""
    del low_bounds, high_bounds, maxpts
    raise NotImplementedError(
        "only 1D box integrations are supported; use `integrate_box_1d`")

  def set_bandwidth(self, bw_method=None):
    """This method is not implemented in the JAX interface."""
    del bw_method
    raise NotImplementedError(
        "dynamically changing the bandwidth method is not supported")

  def _reshape_points(self, points):
    if dtypes.issubdtype(lax.dtype(points), np.complexfloating):
      raise NotImplementedError(
          "gaussian_kde does not support complex coordinates")
    points = jnp.atleast_2d(points)
    d, m = points.shape
    if d != self.d:
      if d == 1 and m == self.d:
        points = jnp.reshape(points, (self.d, 1))
      else:
        raise ValueError(
            "points have dimension {}, dataset has dimension {}".format(
                d, self.d))
    return points

View on GitHub (pinned to 1e1c6a8fc0)

Solutions

1. Create a new gaussian_kde with the new bw_method: kde = gaussian_kde(kde.dataset, bw_method=new_bw, weights=kde.weights)
2. Structure CV loops to construct a fresh KDE per bandwidth candidate
3. Cache datasets/weights and reuse them when re-instantiating to keep the cost low

Example fix

// before
kde.set_bandwidth(0.2)
// after
kde = gaussian_kde(data, bw_method=0.2, weights=w)

try:
    kde.set_bandwidth(bw)
except NotImplementedError:
    kde = gaussian_kde(kde.dataset, bw_method=bw, weights=kde.weights)

When it happens

Trigger: Calling kde.set_bandwidth('silverman') or kde.set_bandwidth(0.3) on a JAX gaussian_kde instance.

Common situations: Porting scipy workflows that retune bandwidth after fitting; cross-validation loops that reuse a fitted KDE and only update bw_method.

Understand the failure class

Background: UnsupportedOperationException and "is not supported" errors: when a library deliberately refuses a call — this error's family across 30 libraries.

Related errors

• only 1D box integrations are supported; use `integrate_box_1
• gaussian_kde does not support complex coordinates
• convolve2d() only supports boundary='fill', fillvalue=0
• correlate2d() only supports boundary='fill', fillvalue=0
• overwrite_data argument not implemented.

AI-assisted analysis of jax-ml/jax@1e1c6a8fc0 (2026-08-27). Data as JSON: /api/errors/55d1dbfcbf7ca000. Report an issue: GitHub.