From 94de39dde2e616a01683c518023d0fab72b4e103 Mon Sep 17 00:00:00 2001
From: 游雁 <zhifu.gzf@alibaba-inc.com>
Date: 星期一, 19 二月 2024 22:21:50 +0800
Subject: [PATCH] aishell example
---
runtime/websocket/bin/funasr-wss-server.cpp | 45 +++++++++++++++++++++++++--------------------
1 files changed, 25 insertions(+), 20 deletions(-)
diff --git a/runtime/websocket/bin/funasr-wss-server.cpp b/runtime/websocket/bin/funasr-wss-server.cpp
index 9c3af70..bff4f66 100644
--- a/runtime/websocket/bin/funasr-wss-server.cpp
+++ b/runtime/websocket/bin/funasr-wss-server.cpp
@@ -26,6 +26,10 @@
}
int main(int argc, char* argv[]) {
+#ifdef _WIN32
+ #include <windows.h>
+ SetConsoleOutputCP(65001);
+#endif
try {
google::InitGoogleLogging(argv[0]);
@@ -91,7 +95,7 @@
"0.0.0.0", "string");
TCLAP::ValueArg<int> port("", "port", "port", false, 10095, "int");
TCLAP::ValueArg<int> io_thread_num("", "io-thread-num", "io thread num",
- false, 8, "int");
+ false, 2, "int");
TCLAP::ValueArg<int> decoder_thread_num(
"", "decoder-thread-num", "decoder thread num", false, 8, "int");
TCLAP::ValueArg<int> model_thread_num("", "model-thread-num",
@@ -111,7 +115,7 @@
TCLAP::ValueArg<std::string> lm_dir("", LM_DIR,
"the LM model path, which contains compiled models: TLG.fst, config.yaml ", false, "damo/speech_ngram_lm_zh-cn-ai-wesp-fst", "string");
TCLAP::ValueArg<std::string> lm_revision(
- "", "lm-revision", "LM model revision", false, "v1.0.1", "string");
+ "", "lm-revision", "LM model revision", false, "v1.0.2", "string");
TCLAP::ValueArg<std::string> hotword("", HOTWORD,
"the hotword file, one hotword perline, Format: Hotword Weight (could be: 闃块噷宸村反 20)",
false, "/workspace/resources/hotwords.txt", "string");
@@ -183,7 +187,7 @@
std::string s_itn_path = model_path[ITN_DIR];
std::string s_lm_path = model_path[LM_DIR];
- std::string python_cmd = "python -m funasr.utils.runtime_sdk_download_tool --type onnx --quantize True ";
+ std::string python_cmd = "python -m funasr.download.runtime_sdk_download_tool --type onnx --quantize True ";
if(vad_dir.isSet() && !s_vad_path.empty()){
std::string python_cmd_vad;
@@ -226,28 +230,29 @@
std::string down_asr_path;
std::string down_asr_model;
+ // modify model-revision by model name
+ size_t found = s_asr_path.find("speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404");
+ if (found != std::string::npos) {
+ model_path["model-revision"]="v1.2.4";
+ }
+
+ found = s_asr_path.find("speech_paraformer-large-contextual_asr_nat-zh-cn-16k-common-vocab8404");
+ if (found != std::string::npos) {
+ model_path["model-revision"]="v1.0.5";
+ }
+
+ found = s_asr_path.find("speech_paraformer-large_asr_nat-en-16k-common-vocab10020");
+ if (found != std::string::npos) {
+ model_path["model-revision"]="v1.0.0";
+ s_itn_path="";
+ s_lm_path="";
+ }
+
if (access(s_asr_path.c_str(), F_OK) == 0){
// local
python_cmd_asr = python_cmd + " --model-name " + s_asr_path + " --export-dir ./ " + " --model_revision " + model_path["model-revision"];
down_asr_path = s_asr_path;
}else{
- size_t found = s_asr_path.find("speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404");
- if (found != std::string::npos) {
- model_path["model-revision"]="v1.2.4";
- }
-
- found = s_asr_path.find("speech_paraformer-large-contextual_asr_nat-zh-cn-16k-common-vocab8404");
- if (found != std::string::npos) {
- model_path["model-revision"]="v1.0.5";
- }
-
- found = s_asr_path.find("speech_paraformer-large_asr_nat-en-16k-common-vocab10020");
- if (found != std::string::npos) {
- model_path["model-revision"]="v1.0.0";
- s_itn_path="";
- s_lm_path="";
- }
-
// modelscope
LOG(INFO) << "Download model: " << s_asr_path << " from modelscope: ";
python_cmd_asr = python_cmd + " --model-name " + s_asr_path + " --export-dir " + s_download_model_dir + " --model_revision " + model_path["model-revision"];
--
Gitblit v1.9.1