make it work on Windows

Author: CodingNinja2020

CMakeLists.txt

@@ -559,6 +559,13 @@ target_link_libraries(engine_runtime PUBLIC ggml)
 target_link_libraries(engine_runtime PRIVATE sentencepiece cjson_vendor yaml_vendor)
 if (ENGINE_ENABLE_OPENMP)
     target_link_libraries(engine_runtime PRIVATE OpenMP::OpenMP_CXX)
+    if (MSVC)
+        # MSVC's default /openmp implements only OpenMP 2.0 and rejects the
+        # '#pragma omp simd' directives in longformer_attention.cpp (error C7660).
+        # /openmp:experimental enables the OpenMP 4.0 SIMD support; it overrides the
+        # /openmp added by OpenMP::OpenMP_CXX above (harmless D9025 override notice).
+        target_compile_options(engine_runtime PRIVATE /openmp:experimental)
+    endif()
 endif()
 
 if (ENGINE_ENABLE_CUDA)

_env.bat

@@ -0,0 +1,47 @@
+@echo off
+REM _env.bat -- shared environment detection for the audio.cpp .bat launchers.
+REM Called (not run) by run_webui.bat / run_server.bat / run_cli_tts.bat. It sets
+REM common variables and deliberately does NOT use setlocal, so they propagate back
+REM to the caller. Change detection logic here only.
+REM
+REM Exports: ROOT BUNDLE WEBUI_DIR PY HAS_CUDA BACKEND SERVER_EXE CLI_EXE GGUF_EXE
+
+REM --- ROOT = this script's directory, without the trailing backslash ---
+set "ROOT=%~dp0"
+if "%ROOT:~-1%"=="\" set "ROOT=%ROOT:~0,-1%"
+
+REM Dev tree: the repo root doubles as the bundle (models\ live under it; the
+REM binaries live under build\). webui.py's own _find_bundle_root handles this.
+set "BUNDLE=%ROOT%"
+set "WEBUI_DIR=%ROOT%\webui"
+
+REM --- Python with the deps (gradio/requests/torch/safetensors/opencc/...) ---
+REM Order: explicit override, project venv (Scripts\ on Windows), then a bundle venv.
+set "PY="
+if defined AUDIOCPP_PYTHON if exist "%AUDIOCPP_PYTHON%" set "PY=%AUDIOCPP_PYTHON%"
+if not defined PY if exist "%ROOT%\venv\Scripts\python.exe" set "PY=%ROOT%\venv\Scripts\python.exe"
+if not defined PY if exist "%ROOT%\venv\python.exe"         set "PY=%ROOT%\venv\python.exe"
+if not defined PY if exist "%BUNDLE%\venv\Scripts\python.exe" set "PY=%BUNDLE%\venv\Scripts\python.exe"
+if not defined PY if exist "%BUNDLE%\venv\python.exe"         set "PY=%BUNDLE%\venv\python.exe"
+
+REM --- CUDA present? (NVIDIA driver installs nvcuda.dll in System32) ---
+set "HAS_CUDA="
+if exist "%SystemRoot%\System32\nvcuda.dll" set "HAS_CUDA=1"
+
+REM --- Locate the from-source binaries. The default Visual Studio generator nests
+REM them in build\bin\Release (multi-config); Ninja/Makefiles use build\bin. ---
+set "BIN="
+if exist "%ROOT%\build\bin\Release\audiocpp_server.exe" set "BIN=%ROOT%\build\bin\Release"
+if not defined BIN if exist "%ROOT%\build\bin\audiocpp_server.exe" set "BIN=%ROOT%\build\bin"
+if defined BIN set "SERVER_EXE=%BIN%\audiocpp_server.exe"
+if defined BIN set "CLI_EXE=%BIN%\audiocpp_cli.exe"
+if defined BIN set "GGUF_EXE=%BIN%\audiocpp_gguf.exe"
+
+REM --- BACKEND: read the actual build's GGML_CUDA flag from its CMakeCache, so we
+REM never advertise a GPU backend a CPU-only build can't serve. cuda when ON, else cpu. ---
+set "BACKEND=cpu"
+if exist "%ROOT%\build\CMakeCache.txt" (
+    for /f "tokens=2 delims==" %%A in ('findstr /b /c:"GGML_CUDA:BOOL" "%ROOT%\build\CMakeCache.txt" 2^>nul') do (
+        if /I "%%A"=="ON" set "BACKEND=cuda"
+    )
+)

tools/model_manager.py

@@ -32,24 +32,41 @@
 yaml: Any = None
 
 
+def _ensure_download_deps() -> None:
+    """Import only what a plain snapshot/file download needs: huggingface_hub + yaml.
+
+    Downloading model weights must not require Torch — importing Torch pulls in
+    native DLLs (c10.dll etc.) that can fail to load in some environments (e.g. a
+    conda-based interpreter on Windows), which would needlessly block downloads for
+    models whose loading/inference happens entirely in the C++ server."""
+    global hf_hub_download, yaml
+    if hf_hub_download is not None:
+        return
+    from huggingface_hub import hf_hub_download as _hf_hub_download
+    import yaml as _yaml
+
+    hf_hub_download = _hf_hub_download
+    yaml = _yaml
+
+
 def _ensure_install_deps() -> None:
-    """Import Torch / huggingface_hub / safetensors / yaml on first install/convert use."""
+    """Import Torch / huggingface_hub / safetensors / yaml on first install/convert use.
+
+    Only needed for conversion (torch.load / safetensors round-trips). Snapshot
+    downloads should call _ensure_download_deps() instead, so they don't drag in Torch."""
     global torch, hf_hub_download, safe_open, load_file, save_file, yaml
     if torch is not None:
         return
+    _ensure_download_deps()
     import torch as _torch
-    from huggingface_hub import hf_hub_download as _hf_hub_download
     from safetensors import safe_open as _safe_open
     from safetensors.torch import load_file as _load_file
     from safetensors.torch import save_file as _save_file
-    import yaml as _yaml
 
     torch = _torch
-    hf_hub_download = _hf_hub_download
     safe_open = _safe_open
     load_file = _load_file
     save_file = _save_file
-    yaml = _yaml
 
 
 REPO_ROOT = Path(__file__).resolve().parents[1]
@@ -2661,7 +2678,6 @@ def command_info(args: argparse.Namespace) -> int:
 
 
 def command_install(args: argparse.Namespace) -> int:
-    _ensure_install_deps()
     package = PACKAGE_BY_ID.get(args.package_id)
     if package is None:
         raise RuntimeError(f"unknown package id: {args.package_id}")
@@ -2671,10 +2687,15 @@ def command_install(args: argparse.Namespace) -> int:
     if isinstance(source, UnsupportedSource):
         raise RuntimeError(f"{package.id} is not installable: {source.reason}")
     if isinstance(source, SnapshotSource):
+        # Plain download: huggingface_hub only, no Torch DLLs.
+        _ensure_download_deps()
         install_path = install_snapshot(package, source, models_root, args.overwrite)
     elif isinstance(source, CompositeSnapshotSource):
+        # Composite snapshots may run a Torch post-process step, so bring in full deps.
+        _ensure_install_deps()
         install_path = install_composite_snapshot(package, source, models_root, args.overwrite)
     else:
+        _ensure_install_deps()
         install_path = install_converter(
             package,
             source,

webui/webui.py

@@ -215,7 +215,22 @@ def _cmake_cache_backend(bin_dir):
     """gpu/cpu for a build tree whose directory name doesn't say, read from its
     CMakeCache.txt (GGML_CUDA:BOOL=ON). Returns None when there is no cache to
     read — an installed tree, or a layout we don't recognize."""
-    cache = os.path.join(os.path.dirname(bin_dir), "CMakeCache.txt")
+    # Walk up from the bin dir to find the build tree's CMakeCache.txt. Single-config
+    # generators put the exe in build/bin (cache one level up); multi-config ones
+    # (Visual Studio) nest it in build/bin/Release, so the cache sits two levels up.
+    cache = None
+    d = os.path.dirname(bin_dir)
+    for _ in range(4):
+        cand = os.path.join(d, "CMakeCache.txt")
+        if os.path.isfile(cand):
+            cache = cand
+            break
+        parent = os.path.dirname(d)
+        if parent == d:
+            break
+        d = parent
+    if cache is None:
+        return None
     try:
         with open(cache, encoding="utf-8", errors="replace") as fh:
             for line in fh:
@@ -240,8 +255,14 @@ def _discover_dev_bin_dirs():
                       if os.path.isfile(os.path.join(d, SERVER_EXE_NAME)))
         if hits:
             out[backend] = hits[-1]
-    plain = os.path.join(PROJECT_ROOT, "build", "bin")
-    if os.path.isfile(os.path.join(plain, SERVER_EXE_NAME)):
+    # A plain `cmake -B build` lands in build/bin with single-config generators
+    # (Ninja, Makefiles); multi-config generators (Visual Studio, Xcode) nest the
+    # exe in a per-config subdir, so also look in build/bin/Release and .../Debug.
+    for plain in (os.path.join(PROJECT_ROOT, "build", "bin"),
+                  os.path.join(PROJECT_ROOT, "build", "bin", "Release"),
+                  os.path.join(PROJECT_ROOT, "build", "bin", "Debug")):
+        if not os.path.isfile(os.path.join(plain, SERVER_EXE_NAME)):
+            continue
         backend = _cmake_cache_backend(plain)
         # Only fill a backend the named-directory scan didn't already find, so an
         # explicit build/linux-cuda-release still wins over a stale plain build/.