chore: take every dependency bump in one pass

[Ayyankhan101]

Supersedes #1 #2 #3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14.

All fourteen Dependabot PRs were based on cb6f48f — the pre-fix master, where
cargo clippy --all-targets --all-features -- -D warnings already exited 101. Every one of
them showed Clippy, MSRV and all three test jobs failing, including #14, which changes
nothing but a GitHub Actions version. Those failures were inherited from the base, not
caused by the bumps, so their CI said nothing useful.

Applying them together against current master and testing locally is both faster and more
meaningful than fourteen rebase-and-wait cycles.

	from	to
candle-core	0.8.4	0.11.0
candle-nn	0.8.4	0.11.0
tokenizers	0.21.4	0.23.1
safetensors	0.4.5	0.8.0
rand	0.8	0.9
rand_distr	0.4	0.5
criterion	0.5	0.8
clap, bytemuck, tokio		latest patch
actions/checkout	4	7
actions/github-script	7	9
actions/upload-artifact	4	7
codecov/codecov-action	4	7

Migrations needed

Three, all mechanical:

• rand 0.9 removed StdRng::from_entropy in favour of from_os_rng, and renamed
  Rng::gen / gen_range to random / random_range.
• safetensors 0.8 takes an owned Option for serialize's metadata argument rather than
  a reference.

candle 0.11 needed no code changes at all — notable given it is the crate everything
here is built on.

Verification

tests/codegen_parity.rs passes unchanged against candle 0.11, so the upgrade does not
move the numbers. Real CodeGen-350M generation still produces correct Python.

102 tests pass
cargo clippy --all-targets --all-features -- -D warnings   exit 0
cargo fmt --all -- --check                                  clean

[git-mind-pr-guardian]

GitMind PR Review

{'pr_url': '#16', 'repo': 'Ayyankhan101/Transformer-In-Rust', 'pr_num': '16', 'title': 'chore: take every dependency bump in one pass', 'author': 'Ayyankhan101', 'additions': 299, 'deletions': 402, 'changed_files': 14, 'review': '## Summary\nAutomatic review could not be generated — please review the diff manually.\n\n## Issues\nNo significant issues found.\n\n## Verdict\n💬 NEEDS DISCUSSION\nThe automated review failed; human review required.'}

---

Auto-generated by GitMind — AI-powered code analysis.