From a05e753d11d9c36983ec4e58c421dbcf86d1dcd4 Mon Sep 17 00:00:00 2001
From: Xian Shi <40013335+R1ckShi@users.noreply.github.com>
Date: 星期二, 17 十月 2023 16:47:27 +0800
Subject: [PATCH] Merge branch 'main' into dev_onnx
---
funasr/runtime/onnxruntime/src/vocab.h | 10 +++++++---
1 files changed, 7 insertions(+), 3 deletions(-)
diff --git a/funasr/runtime/onnxruntime/src/vocab.h b/funasr/runtime/onnxruntime/src/vocab.h
index 6c4e523..23b4bd6 100644
--- a/funasr/runtime/onnxruntime/src/vocab.h
+++ b/funasr/runtime/onnxruntime/src/vocab.h
@@ -5,13 +5,14 @@
#include <stdint.h>
#include <string>
#include <vector>
+#include <map>
using namespace std;
namespace funasr {
class Vocab {
private:
vector<string> vocab;
- bool IsChinese(string ch);
+ std::map<string, int> token_id;
bool IsEnglish(string ch);
void LoadVocabFromYaml(const char* filename);
@@ -19,8 +20,11 @@
Vocab(const char *filename);
~Vocab();
int Size();
- string Vector2String(vector<int> in);
- string Vector2StringV2(vector<int> in);
+ bool IsChinese(string ch);
+ void Vector2String(vector<int> in, std::vector<std::string> &preds);
+ string Vector2StringV2(vector<int> in, std::string language="");
+ string WordFormat(std::string word);
+ int GetIdByToken(const std::string &token);
};
} // namespace funasr
--
Gitblit v1.9.1