Merge branch 'main' of https://github.com/alibaba-damo-academy/FunASR into main

README.md

@@ -13,7 +13,6 @@
 | [**Highlights**](#highlights)
 | [**Installation**](#installation)
 | [**Docs**](https://alibaba-damo-academy.github.io/FunASR/en/index.html)
-| [**Tutorial_CN**](https://github.com/alibaba-damo-academy/FunASR/wiki#funasr%E7%94%A8%E6%88%B7%E6%89%8B%E5%86%8C)
 | [**Papers**](https://github.com/alibaba-damo-academy/FunASR#citations)
 | [**Runtime**](https://github.com/alibaba-damo-academy/FunASR/tree/main/funasr/runtime)
 | [**Model Zoo**](https://github.com/alibaba-damo-academy/FunASR/blob/main/docs/model_zoo/modelscope_models.md)

egs_modelscope/asr/TEMPLATE/README.md

@@ -20,11 +20,13 @@ rec_result = inference_pipeline(audio_in='https://isv-data.oss-cn-hangzhou.aliyu
 print(rec_result)
 ```
 #### [Paraformer-online Model](https://www.modelscope.cn/models/damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-online/summary)
+##### Streaming Decoding
 ```python
 inference_pipeline = pipeline(
     task=Tasks.auto_speech_recognition,
     model='damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-online',
-    model_revision='v1.0.6',
+    model_revision='v1.0.4',
+    update_model='v1.0.4',
     mode='paraformer_streaming'
     )
 import soundfile
@@ -42,6 +44,23 @@ speech_chunk = speech[chunk_stride:chunk_stride+chunk_stride]
 rec_result = inference_pipeline(audio_in=speech_chunk, param_dict=param_dict)
 print(rec_result)
 ```
+
+##### Fake Streaming Decoding
+```python
+from modelscope.pipelines import pipeline
+from modelscope.utils.constant import Tasks
+
+inference_pipeline = pipeline(
+    task=Tasks.auto_speech_recognition,
+    model='damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-online',
+    model_revision='v1.0.6',
+    update_model='v1.0.6',
+    mode="paraformer_fake_streaming"
+)
+audio_in='https://isv-data.oss-cn-hangzhou.aliyuncs.com/ics/MaaS/ASR/test_audio/asr_example_zh.wav'
+rec_result = inference_pipeline(audio_in=audio_in)
+print(rec_result)
+```
 Full code of demo, please ref to [demo](https://github.com/alibaba-damo-academy/FunASR/discussions/241)
 
 #### [UniASR Model](https://www.modelscope.cn/models/damo/speech_UniASR_asr_2pass-zh-cn-8k-common-vocab3445-pytorch-online/summary)

egs_modelscope/asr/paraformer/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-online/demo.py

@@ -5,6 +5,7 @@ inference_pipeline = pipeline(
     task=Tasks.auto_speech_recognition,
     model='damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-online',
     model_revision='v1.0.6',
+    update_model='v1.0.6',
     mode="paraformer_fake_streaming"
 )
 audio_in='https://isv-data.oss-cn-hangzhou.aliyuncs.com/ics/MaaS/ASR/test_audio/asr_example_zh.wav'

egs_modelscope/asr/paraformer/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-online/demo_online.py

@@ -14,7 +14,8 @@ os.environ["MODELSCOPE_CACHE"] = "./"
 inference_pipeline = pipeline(
     task=Tasks.auto_speech_recognition,
     model='damo/speech_paraformer-large_asr_nat-zh-cn-16k-common-vocab8404-online',
-    model_revision='v1.0.6',
+    model_revision='v1.0.4',
+    update_model='v1.0.4',
     mode="paraformer_streaming"
 )
 

egs_modelscope/asr/paraformer/speech_paraformer_asr_nat-zh-cn-16k-common-vocab8404-online/demo.py

@@ -5,6 +5,7 @@ inference_pipeline = pipeline(
     task=Tasks.auto_speech_recognition,
     model='damo/speech_paraformer_asr_nat-zh-cn-16k-common-vocab8404-online',
     model_revision='v1.0.6',
+    update_model='v1.0.6',
     mode="paraformer_fake_streaming"
 )
 audio_in='https://isv-data.oss-cn-hangzhou.aliyuncs.com/ics/MaaS/ASR/test_audio/asr_example_zh.wav'

egs_modelscope/asr/paraformer/speech_paraformer_asr_nat-zh-cn-16k-common-vocab8404-online/demo_online.py

@@ -14,7 +14,8 @@ os.environ["MODELSCOPE_CACHE"] = "./"
 inference_pipeline = pipeline(
     task=Tasks.auto_speech_recognition,
     model='damo/speech_paraformer_asr_nat-zh-cn-16k-common-vocab8404-online',
-    model_revision='v1.0.6',
+    model_revision='v1.0.4',
+    update_model='v1.0.4',
     mode="paraformer_streaming"
 )
 

egs_modelscope/tp/TEMPLATE/README.md

@@ -11,7 +11,7 @@ from modelscope.utils.constant import Tasks
 inference_pipeline = pipeline(
     task=Tasks.speech_timestamp,
     model='damo/speech_timestamp_prediction-v1-16k-offline',
-    output_dir=None)
+    model_revision='v1.1.0')
 
 rec_result = inference_pipeline(
     audio_in='https://isv-data.oss-cn-hangzhou.aliyuncs.com/ics/MaaS/ASR/test_audio/asr_example_timestamps.wav',

fun_text_processing/inverse_text_normalization/id/id_unit_test.tsv

@@ -1,10 +1,10 @@
 dua ribu dua puluh dua	2022
-tiga ribu	300
+tiga ribu	3000
 sembilan ribu sembilan ratus sembilan puluh sembilan	9999
-seribu satu	100001
+seribu satu	1001
-ribu	100
+ribu	1000
 seribu	1000
-seribu dua ratus delapan puluh sembilan	10289
+seribu dua ratus delapan puluh sembilan	1289
 ribu dua ratus delapan puluh sembilan	1289
 nol satu dua tiga empat lima enam tujuh delapan sembilan	01 2345-6789
 empat belas	14
@@ -22,8 +22,8 @@ satu miliar	1 miliar
 seratus dua puluh tiga	123
 ratus dua puluh tiga	123
 dua puluh empat maret 	24 maret
-ribu tujuh puluh enam	10076
+ribu tujuh puluh enam	1076
-seribu tujuh puluh enam	100076
+seribu tujuh puluh enam	1076
-ribu tujuh puluh enam rupiah	10076 rupiah
+ribu tujuh puluh enam rupiah	1076 rupiah
 tujuh puluh enam	76
-ditambah enam dua dua satu enam lima tiga sembilan nol enam nol lima	+62 21 6539-0605
+ditambah enam dua dua satu enam lima tiga sembilan nol enam nol lima	+62 21 6539-0605

fun_text_processing/inverse_text_normalization/id/taggers/cardinal.py

@@ -26,11 +26,10 @@ class CardinalFst(GraphFst):
         graph_teen = pynini.string_file(get_abs_path("data/numbers/teen.tsv"))
         graph_hundreds = pynini.string_file(get_abs_path("data/numbers/hundreds.tsv"))
         graph_thousand = pynini.string_file(get_abs_path("data/numbers/thousand.tsv"))
-
+        
-        graph_cents = pynini.cross("seratus", "100") | pynini.cross("ratus", "100") | pynini.union(graph_hundreds, pynutil.insert("0"))
         graph_hundred = pynini.cross("ratus", "") | pynini.cross("seratus", "")
 
-        graph_hundred_component = pynini.union(graph_digit + delete_space + graph_hundred, pynutil.insert("00"))
+        graph_hundred_component = pynini.union(graph_digit + delete_space + graph_hundred, pynutil.insert("0"))
         graph_hundred_component += delete_space
         graph_hundred_component += pynini.union(
             graph_teen | pynutil.insert("00"),
@@ -44,8 +43,8 @@ class CardinalFst(GraphFst):
                 (graph_ties | pynutil.insert("0")) + delete_space + (
                             graph_digit | pynutil.insert("0")),
         )
-        graph_hundred_component = graph_hundred_component | graph_cents | graph_one_hundred_component
+        graph_hundred_component = graph_hundred_component | graph_one_hundred_component
-
+    
         graph_hundred_component_at_least_one_none_zero_digit = graph_hundred_component @ (
             pynini.closure(DAMO_DIGIT) + (DAMO_DIGIT - "0") + pynini.closure(DAMO_DIGIT)
         )
@@ -54,14 +53,12 @@ class CardinalFst(GraphFst):
         )
         graph_thousand = pynini.cross("ribu", "") | pynini.cross("seribu", "")
         graph_one_thousand_component = pynini.union(pynini.cross("ribu", "1") | pynini.cross("seribu", "1"))
-        graph_thousand_cents = pynini.cross("seribu", "10") | pynini.cross("ribu","10") | pynini.union(graph_thousand, pynutil.insert(""))
+       
         graph_thousands = pynini.union(
             graph_hundred_component_at_least_one_none_zero_digit + delete_space + (pynutil.delete("ribu") | pynutil.delete("seribu")),
             pynutil.insert("000", weight=0.1),
         )
-        graph_thousand_component = pynini.union(graph_digit + delete_space + graph_thousand, pynutil.insert("000"))
+        graph_thousands = graph_thousands | (pynutil.insert("00") + graph_one_thousand_component)
-        graph_thousand_component += delete_space
-        graph_thousands = graph_thousands | graph_thousand_cents | graph_thousand_component | graph_one_thousand_component
 
         graph_million = pynini.union(
             graph_hundred_component_at_least_one_none_zero_digit + delete_space + (pynutil.delete("juta") | pynutil.delete("sejuta")),

funasr/runtime/html5/static/main.js

@@ -145,7 +145,9 @@ function stop() {
 	isRec = false;
     info_div.innerHTML="请等候...";
 	btnStop.disabled = true;
-	setTimeout(function(){btnStart.disabled = false;info_div.innerHTML="请点击开始";}, 3000 );
+	setTimeout(function(){
+		console.log("call stop ws!");
+		wsconnecter.wsStop();btnStart.disabled = false;info_div.innerHTML="请点击开始";}, 3000 );
 	rec.stop(function(blob,duration){
   
 		console.log(blob);

funasr/runtime/html5/static/wsconnecter.js

@@ -28,7 +28,11 @@ function WebSocketConnectMethod( config ) { //定义socket连接方法类
 		if ( 'WebSocket' in window ) {
 			speechSokt = new WebSocket( Uri ); // 定义socket连接对象
 			speechSokt.onopen = function(e){onOpen(e);}; // 定义响应函数
-			speechSokt.onclose = function(e){onClose(e);};
+			speechSokt.onclose = function(e){
+			    console.log("onclose ws!");
+			    speechSokt.close();
+				onClose(e);
+				};
 			speechSokt.onmessage = function(e){onMessage(e);};
 			speechSokt.onerror = function(e){onError(e);};
 			return 1;
@@ -42,6 +46,7 @@ function WebSocketConnectMethod( config ) { //定义socket连接方法类
 	// 定义停止与发送函数
 	this.wsStop = function () {
 		if(speechSokt != undefined) {
+			console.log("stop ws!");
 			speechSokt.close();
 		}
 	};

funasr/runtime/python/websocket/wss_srv_asr.py

@@ -58,16 +58,36 @@ inference_pipeline_asr_online = pipeline(
     model=args.asr_model_online,
     ngpu=args.ngpu,
     ncpu=args.ncpu,
-    model_revision='v1.0.6',
+    model_revision='v1.0.4',
+    update_model='v1.0.4',
     mode='paraformer_streaming')
 
-print("model loaded")
+print("model loaded! only support one client at the same time now!!!!")
 
+async def ws_reset(websocket):
+    print("ws reset now, total num is ",len(websocket_users))
+    websocket.param_dict_asr_online = {"cache": dict()}
+    websocket.param_dict_vad = {'in_cache': dict(), "is_final": True}
+    websocket.param_dict_asr_online["is_final"]=True
+    audio_in=b''.join(np.zeros(int(16000),dtype=np.int16))
+    inference_pipeline_vad(audio_in=audio_in, param_dict=websocket.param_dict_vad)
+    inference_pipeline_asr_online(audio_in=audio_in, param_dict=websocket.param_dict_asr_online)
+    await websocket.close()
+    
+    
+async def clear_websocket():
+   for websocket in websocket_users:
+       await ws_reset(websocket)
+   websocket_users.clear()
+ 
+ 
+       
 async def ws_serve(websocket, path):
     frames = []
     frames_asr = []
     frames_asr_online = []
     global websocket_users
+    await clear_websocket()
     websocket_users.add(websocket)
     websocket.param_dict_asr = {}
     websocket.param_dict_asr_online = {"cache": dict()}
@@ -139,7 +159,8 @@ async def ws_serve(websocket, path):
 
      
     except websockets.ConnectionClosed:
-        print("ConnectionClosed...", websocket_users)
+        print("ConnectionClosed...", websocket_users,flush=True)
+        await ws_reset(websocket)
         websocket_users.remove(websocket)
     except websockets.InvalidState:
         print("InvalidState...")