chore(llama.cpp): detect version and bump to b3995 for win (#3447)

* chore(llama.cpp): auto detech llama.cpp version and bump to b3995 for win

* ci(build): fetch tags before getting llama.cpp version

* ci(build): do not display all fetched tags when build

.github/workflows/release.yml

@@ -249,14 +249,14 @@ jobs:
 
   package-win-cuda:
     runs-on: ubuntu-latest
-    env:
-      LLAMA_CPP_VERSION: b3571
     needs: [release-binary]
     permissions:
       contents: write
     steps:
       - name: Checkout
         uses: actions/checkout@v3
+        with:
+          submodules: recursive
 
       - name: Download all artifacts
         uses: actions/download-artifact@v3

ci/package-win.sh

@@ -1,8 +1,12 @@
 #!/bin/bash
 
+# get current bash file directory
+PROJECT_ROOT="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
+LLAMA_CPP_PATH="${PROJECT_ROOT}/crates/llama-cpp-server/llama.cpp"
+
 # Input variables
-TABBY_VERSION=${TABBY_VERSION:-dev}
-LLAMA_CPP_VERSION=${LLAMA_CPP_VERSION:-b3571}
+LLAMA_CPP_VERSION=${LLAMA_CPP_VERSION:-$(cd ${LLAMA_CPP_PATH} && git fetch --tags origin >/dev/null && git describe --tags --abbrev=0)}
+echo "LLAMA_CPP_VERSION=${LLAMA_CPP_VERSION}"
 LLAMA_CPP_PLATFORM=${LLAMA_CPP_PLATFORM:-cuda-cu11.7.1-x64}
 OUTPUT_NAME=${OUTPUT_NAME:-tabby_x86_64-windows-msvc-cuda117}