fix(studio): pin llama.cpp to b8637 release (Gemma 4 support) (#4796)

ggml-org/llama.cpp b8637 includes Gemma 4 support (ggml-org/llama.cpp#21309).
Revert the temporary "master" default back to a pinned release tag.

This eliminates the HTTP 422 errors from the prebuilt resolver (which
could not find a release matching "master"), avoids unnecessary source
builds, and restores prebuilt binary downloads on all platforms.

Co-authored-by: Daniel Han <danielhanchen@users.noreply.github.com>

studio/install_llama_prebuilt.py

@@ -60,7 +60,7 @@ def env_int(name: str, default: int, *, minimum: int | None = None) -> int:
     return value
 
 
-DEFAULT_LLAMA_TAG = os.environ.get("UNSLOTH_LLAMA_TAG", "master")
+DEFAULT_LLAMA_TAG = os.environ.get("UNSLOTH_LLAMA_TAG", "b8637")
 # Force all installs to use mainline llama.cpp from ggml-org.
 # Previously: DEFAULT_PUBLISHED_REPO = os.environ.get("UNSLOTH_LLAMA_RELEASE_REPO", "unslothai/llama.cpp")
 DEFAULT_PUBLISHED_REPO = "ggml-org/llama.cpp"

studio/setup.ps1

@@ -29,7 +29,7 @@ $PackageDir = Split-Path -Parent $ScriptDir
 # --------------------------------------------------------------------------
 $DefaultLlamaPrForce = ""
 $DefaultLlamaSource = "https://github.com/ggml-org/llama.cpp"
-$DefaultLlamaTag = "master"
+$DefaultLlamaTag = "b8637"
 
 # Verbose can be enabled either by CLI flag or by UNSLOTH_VERBOSE=1.
 $script:UnslothVerbose = ($env:UNSLOTH_VERBOSE -eq '1')

studio/setup.sh

@@ -19,7 +19,7 @@ RULE=$(printf '\342\224\200%.0s' {1..52})
 # ──────────────────────────────────────────────────────────────────────────
 _DEFAULT_LLAMA_PR_FORCE=""
 _DEFAULT_LLAMA_SOURCE="https://github.com/ggml-org/llama.cpp"
-_DEFAULT_LLAMA_TAG="master"
+_DEFAULT_LLAMA_TAG="b8637"
 
 # ── Colors (same palette as startup_banner / install_python_stack) ──
 if [ -n "${NO_COLOR:-}" ]; then