Update llama.cpp models and Gemma 4 chat templates

[addianto]

No description provided.

[Copilot]

Pull request overview

Updates local llama.cpp presets to newer Gemma 4 / Qwen model entries and refreshes the vendored Gemma 4 chat templates, plus adds an idle-sleep flag to the llama-server start scripts.

Changes:

• Update model preset .ini files (Gemma 4 repo entries, speculative drafting settings, and embedding/reranker model selections).
• Vendor updated “canonical” Gemma 4 chat templates with revised tool-calling / turn-closure logic.
• Add --sleep-idle-seconds to both the PowerShell and POSIX launcher scripts to reduce idle resource usage.

Reviewed changes

Copilot reviewed 5 out of 7 changed files in this pull request and generated 6 comments.

Show a summary per file

File	Description
config/llama.cpp/rinze.ini	Updates Gemma 4 E4B preset, adds speculative drafting config, and refreshes Qwen/embedding/reranker entries.
config/llama.cpp/alice.ini	Updates Gemma 4 presets, increases context size for 26B preset, adds speculative drafting config, and refreshes Qwen/embedding/reranker entries.
config/llama.cpp/gemma-4-e4b-chat_template.jinja	Updates Gemma 4 E4B chat template logic around tool-calls, thinking gating, and turn continuation.
config/llama.cpp/gemma-4-e2b-chat_template.jinja	Updates Gemma 4 E2B chat template logic around tool-calls, thinking gating, and turn continuation.
config/llama.cpp/gemma-4-26b-a4b-chat_template.jinja	Updates Gemma 4 26B A4B chat template logic around tool-calls, thinking gating, and turn continuation.
bin/Start-LlamaServer.ps1	Adds an idle sleep duration and passes --sleep-idle-seconds to llama-server.
bin/start-llama-server.sh	Adds an idle sleep duration and passes --sleep-idle-seconds to llama-server.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.