Add SHL provider.

CMakeLists.txt

@@ -27,6 +27,7 @@ option(SHERPA_ONNX_ENABLE_WASM_NODEJS "Whether to enable WASM for NodeJS" OFF)
 option(SHERPA_ONNX_ENABLE_BINARY "Whether to build binaries" ON)
 option(SHERPA_ONNX_LINK_LIBSTDCPP_STATICALLY "True to link libstdc++ statically. Used only when BUILD_SHARED_LIBS is OFF on Linux" ON)
 option(SHERPA_ONNX_USE_PRE_INSTALLED_ONNXRUNTIME_IF_AVAILABLE "True to use pre-installed onnxruntime if available" ON)
+option(SHERPA_ONNX_ENABLE_SHL "Whether to enable SHL for RISC-V" OFF)
 
 set(CMAKE_ARCHIVE_OUTPUT_DIRECTORY "${CMAKE_BINARY_DIR}/lib")
 set(CMAKE_LIBRARY_OUTPUT_DIRECTORY "${CMAKE_BINARY_DIR}/lib")
@@ -112,6 +113,7 @@ message(STATUS "SHERPA_ONNX_ENABLE_WASM_TTS ${SHERPA_ONNX_ENABLE_WASM_TTS}")
 message(STATUS "SHERPA_ONNX_ENABLE_WASM_ASR ${SHERPA_ONNX_ENABLE_WASM_ASR}")
 message(STATUS "SHERPA_ONNX_ENABLE_WASM_NODEJS ${SHERPA_ONNX_ENABLE_WASM_NODEJS}")
 message(STATUS "SHERPA_ONNX_USE_PRE_INSTALLED_ONNXRUNTIME_IF_AVAILABLE ${SHERPA_ONNX_USE_PRE_INSTALLED_ONNXRUNTIME_IF_AVAILABLE}")
+message(STATUS "SHERPA_ONNX_ENABLE_SHL ${SHERPA_ONNX_ENABLE_SHL}")
 
 if(SHERPA_ONNX_ENABLE_WASM_TTS)
   if(NOT SHERPA_ONNX_ENABLE_WASM)
@@ -135,6 +137,10 @@ if(SHERPA_ONNX_ENABLE_WASM)
   add_definitions(-DSHERPA_ONNX_ENABLE_WASM=1)
 endif()
 
+if(SHERPA_ONNX_ENABLE_SHL)
+  add_definitions(-DSHERPA_ONNX_ENABLE_SHL=1)
+endif()
+
 if(NOT CMAKE_CXX_STANDARD)
   set(CMAKE_CXX_STANDARD 14 CACHE STRING "The C++ version to be used.")
 endif()

build-riscv64-linux-gnu.sh

@@ -62,6 +62,7 @@ cmake \
   -DSHERPA_ONNX_ENABLE_JNI=OFF \
   -DSHERPA_ONNX_ENABLE_C_API=OFF \
   -DSHERPA_ONNX_ENABLE_WEBSOCKET=OFF \
+  -DSHERPA_ONNX_ENABLE_SHL=ON \
   -DCMAKE_TOOLCHAIN_FILE=../toolchains/riscv64-linux-gnu.toolchain.cmake \
   ..
 

sherpa-onnx/csrc/provider.cc

@@ -24,6 +24,8 @@ Provider StringToProvider(std::string s) {
     return Provider::kXnnpack;
   } else if (s == "nnapi") {
     return Provider::kNNAPI;
+  } else if (s == "shl") {
+    return Provider::kShl;
   } else {
     SHERPA_ONNX_LOGE("Unsupported string: %s. Fallback to cpu", s.c_str());
     return Provider::kCPU;

sherpa-onnx/csrc/provider.h

@@ -18,6 +18,7 @@ enum class Provider {
   kCoreML = 2,   // CoreMLExecutionProvider
   kXnnpack = 3,  // XnnpackExecutionProvider
   kNNAPI = 4,    // NnapiExecutionProvider
+  kShl = 5,      // kShlExecutionProvider
 };
 
 /**

sherpa-onnx/csrc/session.cc

@@ -9,6 +9,7 @@
 #include <utility>
 #include <vector>
 
+#include "onnxruntime_c_api.h"
 #include "sherpa-onnx/csrc/macros.h"
 #include "sherpa-onnx/csrc/provider.h"
 #if defined(__APPLE__)
@@ -115,6 +116,37 @@ static Ort::SessionOptions GetSessionOptionsImpl(int32_t num_threads,
 #endif
       break;
     }
+
+    case Provider::kShl: {
+#if defined(SHERPA_ONNX_ENABLE_SHL)
+      if (std::find(available_providers.begin(), available_providers.end(),
+                    "ShlExecutionProvider") != available_providers.end()) {
+        // sess_opts.AppendExecutionProvider_SHL({});
+
+        const auto &api = Ort::GetApi();
+        OrtStatus *status =
+            api.OrtSessionOptionsAppendExecutionProvider_Shl(sess_opts, "");
+
+        if (status) {
+          const char *msg = api.GetErrorMessage(status);
+          SHERPA_ONNX_LOGE(
+              "Failed to enable Shl: %s. Available providers: %s. Fallback "
+              "to "
+              "cpu",
+              msg, os.str().c_str());
+          api.ReleaseStatus(status);
+        } else {
+          SHERPA_ONNX_LOGE("Use Shl");
+        }
+      } else
+#endif
+      {
+        SHERPA_ONNX_LOGE("Available providers: %s. Fallback to cpu!",
+                         os.str().c_str());
+      }
+
+      break;
+    }
   }
 
   return sess_opts;