fix(neuron): derive Clone on ImageInput for the CUDA vision dispatch

CUDA type-check in CI failed on commit 24968e9 with E0308:

  error[E0308]: mismatched types
      --> crates/neuron/src/harness/candle.rs:1707:33
   1707 |                                 images.clone(),
        |                                 ^^^^^^^^^^^^^^ expected `Vec<ImageInput>`,
                                                          found `&Vec<ImageInput>`

In Stage B5 the cuda branch of `chat_completion` matches
`&vision_route` to keep the `vision_route: Option<...>` alive for
both arms, which makes `images` bind as `&Vec<ImageInput>`. The
subsequent `images.clone()` call doesn't deep-clone because
`ImageInput` doesn't derive `Clone` — rustc falls back to cloning
the `&Vec` reference, which has the wrong type for the worker job.

The CPU build (non-cuda) compiled fine because that branch is
behind `#[cfg(feature = "cuda")]`; the cuda-check job is what
catches the regression.

Fix: derive `Clone` on `ImageInput`. The clone cost is one
pixel-buffer memcpy per image (~2.4 MiB at fixed 448×448), which
is fine on the chat-completion hot path — vision requests are
rare per second relative to text-only decode.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

crates/neuron/src/harness/device_worker/jobs.rs

@@ -32,6 +32,13 @@ pub struct TpHandle(pub u64);
 /// `Job::EncodeImage`. Pixels are row-major `(c, h, w)` f32 — the
 /// shape `harness::preprocess::preprocess` produces. Carries the
 /// shape inline since `Vec<f32>` is rank-1.
+///
+/// `Clone` so the vision-aware dispatch in `chat_completion` can
+/// match `&vision_route` (carrying borrowed images) and still hand
+/// owned `Vec<ImageInput>` to the worker job. The clone cost is one
+/// pixel-buffer memcpy per image — fine at fixed-resolution sizes
+/// (3 × 448 × 448 × 4 bytes = ~2.4 MiB per image).
+#[derive(Clone)]
 pub struct ImageInput {
     pub pixels: Vec<f32>,
     pub c: usize,