build(deps): Bump oxmera from 0.4.0 to 0.5.2

[dependabot]

Bumps oxmera from 0.4.0 to 0.5.2.

Release notes

Sourced from oxmera's releases.

v0.5.2

What's Changed

• oxmera 0.5.2 — the autograd round by @​vyncint in vyncint/oxmera#102

Full Changelog: vyncint/oxmera@v0.5.1...v0.5.2

v0.5.1

What's Changed

• oxmera 0.5.1 — ten defects found by deep-testing the published 0.5.0 by @​vyncint in vyncint/oxmera#94

Full Changelog: vyncint/oxmera@v0.5.0...v0.5.1

v0.5.0

What's Changed

• test: termlens 0.10.1, and the invariant the PTY suite rests on by @​vyncint in vyncint/oxmera#42
• chore(deps): termlens 0.11, the stability candidate by @​vyncint in vyncint/oxmera#43
• ci: the DCO squash-rewrite equivalence runs both ways by @​vyncint in vyncint/oxmera#44
• ci: termlens 0.11's two stale pins, and the check that missed them by @​vyncint in vyncint/oxmera#77
• fix(nn): prevent neural network layer parameter aliasing by @​mikemikimike in vyncint/oxmera#79
• docs: correct backend and optimizer summary by @​nightcityblade in vyncint/oxmera#78
• docs: refresh oxmera-core status by @​nightcityblade in vyncint/oxmera#80
• ci: the DCO rewrite exemption keys on the committer, not the address shape by @​vyncint in vyncint/oxmera#81
• oxmera 0.5.0 — production hardening (25 issues) by @​vyncint in vyncint/oxmera#83

New Contributors

• @​mikemikimike made their first contribution in vyncint/oxmera#79
• @​nightcityblade made their first contribution in vyncint/oxmera#78

Full Changelog: vyncint/oxmera@v0.4.0...v0.5.0

Changelog

Sourced from oxmera's changelog.

[0.5.2] — 2026-09-16

A second deep test of the published crates, this time reaching the autograd surface the first round never got to. No public API changed.

Fixed

• A deep tape aborted the process. The tape is a linked structure and its derived drop recursed once per node, overflowing the stack at roughly 20k operations in debug and 65k in release — an abort, which no caller can catch. AutogradMeta now dismantles the graph with an explicit worklist, so depth costs heap instead of stack; a 200,000-node tape builds and drops cleanly. Unrolled RNNs and long chains built before backward are the ordinary way to hit this (#96).
• matmul disagreed with itself across its row paths. The single-row path skipped the multiply when the left operand was 0.0, so 0 × ∞ never happened there while the four-row block produced NaN — one call returned [NaN × 8, 0.0, 0.0] for uniform input. Both paths now do the same arithmetic (#98).
• max/min reductions silently discarded NaN. f32::max returns the non-NaN operand, so a NaN that sum propagates and argmax refuses vanished through max — the op on the common path through pooling and attention. NaN now propagates, matching sum (#97).
• Degenerate inputs panicked, aborted, or were silently accepted where a typed error is promised: try_zeros/try_ones/try_full aborted the process on an allocation the allocator refused (they now use try_reserve_exact, which is the entire point of a try_ constructor); eye(n) wrapped n * n and surfaced as an index-out-of-bounds (now a checked, documented panic); pad_dim panicked on an out-of-range dim and accepted it silently when the padding was zero, where its sibling cat has always reported it; kaiming_uniform/xavier_uniform panicked inside rand on a zero fan; and slice turned a reversed range into an empty view (#99).
• Errors named the wrong thing. trace reported an error naming diag, an op the caller never wrote; index_select reported a negative index as index [0], an index that is valid, sending the reader to the wrong element; and matmul's rank error claimed "supported ranks are 2 and 3" although rank-4 and rank-5 both work (#100).
• Documentation described behaviour the code does not have. oxmera-tensor's crate doc and register_backend still taught load-time registration, which 0.5.0 removed — following them produced a BackendUnavailable the docs called impossible. check_param_dtype promised DTypeMismatch and returns InvalidArgument. The README called reshape zero-copy unconditionally. The top-level --help omitted --json and --seed, which the subcommands' own help documents. And no crate carried [package.metadata.docs.rs], so docs.rs rendered oxmera-metal as its non-macOS stub (#101).

Documented

... (truncated)

Commits

• bb5c56c docs: finish the 0.5.0 registration change, correct four claims, cut 0.5.2
• e3e67bc fix(nn): pad_dim reports a bad dim, and a zero fan does not panic the init
• d55aee3 fix(tensor): iterative tape drop, consistent matmul, NaN in reductions, typed...
• 1fa0359 docs: correct the post-0.5.0 claims, and cut 0.5.1
• 53ab86a fix(release): ship the licence texts with every crate
• 5fa82ec fix(nn): loading weights is all-or-nothing, and duplicate names are typed
• c48555e fix(optim): per-parameter bias correction, and state that grows with a group
• 7ae3dc4 fix(tensor): tie gradients, overflow-safe bounds, empty matmul, f64 mean
• 5e67550 docs: STABILITY.md and single-GPU limits; cut the 0.5.0 CHANGELOG and bump
• 0a642d9 ci: semver-checks, release gates, scheduled advisories, and crate metadata
• Additional commits viewable in compare view

[vyncint]

@dependabot rebase