]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
docs : update speculative decoding parameters after refactor (#22397) (#22539)
authorGeorgi Gerganov <redacted>
Mon, 4 May 2026 05:52:07 +0000 (08:52 +0300)
committerGitHub <redacted>
Mon, 4 May 2026 05:52:07 +0000 (08:52 +0300)
commit846262d7875dcabf502a150fa3d7b9c770dde7eb
tree922d145614fbd76e1e51370b3cd2063eacd636be
parent6dcd824fce51e4b03ba2064f83990ff5fee85aeb
docs : update speculative decoding parameters after refactor (#22397) (#22539)

* docs : update speculative decoding parameters after refactor (#22397)

Update docs/speculative.md to reflect the new parameter naming scheme
introduced in PR #22397:

- Replace --draft-max/--draft-min with --spec-draft-n-max/--spec-draft-n-min
- Replace --spec-ngram-size-n/m with per-implementation variants
- Add documentation for all new --spec-ngram-*- parameters
- Update all example commands

Assisted-by: llama.cpp:local pi
* pi : add rule to use gh CLI for GitHub resources

Assisted-by: llama.cpp:local pi
* docs : run llama-gen-docs

* arg : fix typo
.pi/gg/SYSTEM.md
common/arg.cpp
docs/speculative.md
tools/cli/README.md
tools/completion/README.md
tools/server/README.md