add ASR_Model Select

tools/asr/config.py

@@ -0,0 +1,31 @@
+import os
+
+def check_fw_local_models():
+    '''
+    启动时检查本地是否有 Faster Whisper 模型.
+    '''
+    model_size_list = [
+        "tiny",     "tiny.en", 
+        "base",     "base.en", 
+        "small",    "small.en", 
+        "medium",   "medium.en", 
+        "large",    "large-v1", 
+        "large-v2", "large-v3"]
+    for i, size in enumerate(model_size_list):
+        if os.path.exists(f'tools/asr/models/faster-whisper-{size}'):
+            model_size_list[i] = size + '(local)'
+    return model_size_list
+
+asr_dict = {
+    "达摩 ASR (中文)": {
+        'lang': ['zh'],
+        'size': ['large'],
+        'path': 'funasr_asr.py',
+    },
+    "Faster Whisper (多语种)": {
+        'lang': ['auto', 'zh', 'en', 'ja'],
+        'size': check_fw_local_models(),
+        'path': 'fasterwhisper_asr.py'
+    }
+}
+

tools/asr/fasterwhisper_asr.py

@@ -0,0 +1,97 @@
+import argparse
+import os
+import traceback
+import requests
+from glob import glob
+
+from faster_whisper import WhisperModel
+from tqdm import tqdm
+
+from config import check_fw_local_models
+
+os.environ["KMP_DUPLICATE_LIB_OK"]="TRUE"
+
+language_code_list = [
+    "af", "am", "ar", "as", "az", 
+    "ba", "be", "bg", "bn", "bo", 
+    "br", "bs", "ca", "cs", "cy", 
+    "da", "de", "el", "en", "es", 
+    "et", "eu", "fa", "fi", "fo", 
+    "fr", "gl", "gu", "ha", "haw", 
+    "he", "hi", "hr", "ht", "hu", 
+    "hy", "id", "is", "it", "ja", 
+    "jw", "ka", "kk", "km", "kn", 
+    "ko", "la", "lb", "ln", "lo", 
+    "lt", "lv", "mg", "mi", "mk", 
+    "ml", "mn", "mr", "ms", "mt", 
+    "my", "ne", "nl", "nn", "no", 
+    "oc", "pa", "pl", "ps", "pt", 
+    "ro", "ru", "sa", "sd", "si", 
+    "sk", "sl", "sn", "so", "sq", 
+    "sr", "su", "sv", "sw", "ta", 
+    "te", "tg", "th", "tk", "tl", 
+    "tr", "tt", "uk", "ur", "uz", 
+    "vi", "yi", "yo", "zh", "yue",
+    "auto"]
+
+def execute_asr(input_folder, output_folder, model_size, language):
+    if 'local' in model_size:
+        model_size = model_size.split('(')[0]
+        model_path = f'tools/asr/models/faster-whisper-{model_size}'
+    else:
+        model_path = model_size
+    if language == 'auto':
+        language = None #不设置语种由模型自动输出概率最高的语种
+
+    try:
+        model = WhisperModel(model_path, device="cuda", compute_type="float16")
+    except:
+        return print(traceback.format_exc())
+
+    output = []
+    output_file_name = os.path.basename(input_folder)
+    output_file_path = os.path.abspath(f'{output_folder}/{output_file_name}.list')
+
+    if not os.path.exists(output_folder):
+        os.makedirs(output_folder)
+
+    for file in tqdm(glob(os.path.join(input_folder, '**/*.wav'), recursive=True)):
+        try:
+            segments, info = model.transcribe(
+                audio          = file,
+                beam_size      = 5,
+                vad_filter     = True,
+                vad_parameters = dict(min_silence_duration_ms=700),
+                language       = language)
+            text = ''
+            for segment in segments:
+                text += segment.text
+            output.append(f"{file}|{output_file_name}|{info.language.upper()}|{text}")
+        except:
+            return print(traceback.format_exc())
+        
+    with open(output_file_path, "w", encoding="utf-8") as f:
+        f.write("\n".join(output))
+        print(f"ASR 任务完成->标注文件路径: {output_file_path}\n")
+    return output_file_path
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser()
+    parser.add_argument("-i", "--input_folder", type=str, required=True,
+                        help="Path to the folder containing WAV files.")
+    parser.add_argument("-o", "--output_folder", type=str, required=True, 
+                        help="Output folder to store transcriptions.")
+    parser.add_argument("-s", "--model_size", type=str, default='large-v3', 
+                        choices=check_fw_local_models(),
+                        help="Model Size of Faster Whisper")
+    parser.add_argument("-l", "--language", type=str, default='zh', 
+                        choices=language_code_list,
+                        help="Language of the audio files.")
+
+    cmd = parser.parse_args()
+    output_file_path = execute_asr(
+        input_folder  = cmd.input_folder,
+        output_folder = cmd.output_folder,
+        model_size    = cmd.model_size,
+        language      = cmd.language,
+    )

tools/asr/funasr_asr.py

@@ -0,0 +1,66 @@
+# -*- coding:utf-8 -*-
+
+import argparse
+import os
+import traceback
+from tqdm import tqdm
+
+from funasr import AutoModel
+
+path_asr  = 'tools/damo_asr/models/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch'
+path_vad  = 'tools/damo_asr/models/speech_fsmn_vad_zh-cn-16k-common-pytorch'
+path_punc = 'tools/damo_asr/models/punc_ct-transformer_zh-cn-common-vocab272727-pytorch'
+path_asr  = path_asr  if os.path.exists(path_asr)  else "iic/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-pytorch"
+path_vad  = path_vad  if os.path.exists(path_vad)  else "iic/speech_fsmn_vad_zh-cn-16k-common-pytorch"
+path_punc = path_punc if os.path.exists(path_punc) else "iic/punc_ct-transformer_zh-cn-common-vocab272727-pytorch"
+
+model = AutoModel(
+    model               = path_asr,
+    model_revision      = "v2.0.4",
+    vad_model           = path_vad,
+    vad_model_revision  = "v2.0.4",
+    punc_model          = path_punc,
+    punc_model_revision = "v2.0.4",
+)
+
+def execute_asr(input_folder, output_folder, model_size, language):
+    input_file_names = os.listdir(input_folder)
+    input_file_names.sort()
+    
+    output = []
+    output_file_name = os.path.basename(input_folder)
+
+    for name in tqdm(input_file_names):
+        try:
+            text = model.generate(input="%s/%s"%(input_folder, name))[0]["text"]
+            output.append(f"{input_folder}/{name}|{output_file_name}|{language.upper()}|{text}")
+        except:
+            return print(traceback.format_exc())
+
+    output_folder = output_folder or "output/asr_opt"
+    os.makedirs(output_folder, exist_ok=True)
+    output_file_path = os.path.abspath(f'{output_folder}/{output_file_name}.list')
+
+    with open(output_file_path, "w", encoding="utf-8") as f:
+        f.write("\n".join(output))
+        print(f"ASR 任务完成->标注文件路径: {output_file_path}\n")
+    return output_file_path
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser()
+    parser.add_argument("-i", "--input_folder", type=str, required=True,
+                        help="Path to the folder containing WAV files.")
+    parser.add_argument("-o", "--output_folder", type=str, required=True, 
+                        help="Output folder to store transcriptions.")
+    parser.add_argument("-s", "--model_size", type=str, default='large',
+                        help="Model Size of FunASR is Large")
+    parser.add_argument("-l", "--language", type=str, default='zh', choices=['zh'],
+                        help="Language of the audio files.")
+
+    cmd = parser.parse_args()
+    execute_asr(
+        input_folder  = cmd.input_folder,
+        output_folder = cmd.output_folder,
+        model_size    = cmd.model_size,
+        language      = cmd.language,
+    )

tools/asr/models/.gitignore

@@ -0,0 +1,2 @@
+*
+!.gitignore