From caac4ac47ba92b6323d0546e7496045bb72a7144 Mon Sep 17 00:00:00 2001 From: makaveli10 Date: Mon, 31 Jul 2023 14:09:39 +0800 Subject: [PATCH 1/4] return lang, langprrob from transcriber --- transcriber.py | 1 + 1 file changed, 1 insertion(+) diff --git a/transcriber.py b/transcriber.py index e4404e3..ec6e28c 100644 --- a/transcriber.py +++ b/transcriber.py @@ -288,6 +288,7 @@ class WhisperModel: language, language_probability, ) + return language, language_probability else: language_probability = 1 From f20c4153d430b58a36f2d293fcbe8c76d8ff2f3d Mon Sep 17 00:00:00 2001 From: makaveli10 Date: Mon, 31 Jul 2023 14:40:38 +0800 Subject: [PATCH 2/4] set lang to en if not multilingual --- client.py | 2 +- server.py | 3 ++- 2 files changed, 3 insertions(+), 2 deletions(-) diff --git a/client.py b/client.py index f912beb..44fe89c 100644 --- a/client.py +++ b/client.py @@ -251,7 +251,7 @@ if __name__=="__main__": opt = parser.parse_args() print(opt) multilingual=opt.multilingual, - language = opt.language, + language = opt.language if opt.multilingual else "en", task = opt.task c = Client(host=opt.host, port=opt.port) diff --git a/server.py b/server.py index 9879b9f..68c21a1 100644 --- a/server.py +++ b/server.py @@ -28,6 +28,7 @@ def recv_audio(websocket): global clients options = websocket.recv() options = json.loads(options) + print(options) client = ServeClient( websocket, multilingual=options["multilingual"], @@ -55,7 +56,7 @@ class ServeClient: def __init__(self, websocket, task="transcribe", device=None, multilingual=False, language=None): self.data = b"" self.frames = b"" - self.language = language + self.language = language if multilingual else "en" self.task = task self.transcriber = WhisperModel( "small" if multilingual else "small.en", From 6a78c51ce4682b699e0513252c8e0055fdf51d82 Mon Sep 17 00:00:00 2001 From: makaveli10 Date: Mon, 31 Jul 2023 14:42:36 +0800 Subject: [PATCH 3/4] remove debug print --- server.py | 1 - 1 file changed, 1 deletion(-) diff --git a/server.py b/server.py index 68c21a1..d675ef1 100644 --- a/server.py +++ b/server.py @@ -28,7 +28,6 @@ def recv_audio(websocket): global clients options = websocket.recv() options = json.loads(options) - print(options) client = ServeClient( websocket, multilingual=options["multilingual"], From 42c22787130b10636277de0e898370be89aeab13 Mon Sep 17 00:00:00 2001 From: makaveli10 Date: Mon, 31 Jul 2023 12:14:41 +0530 Subject: [PATCH 4/4] language to en if not multilingual --- Audio-Transcription-Chrome/options.js | 3 +++ Audio-Transcription-Firefox/content.js | 3 +++ 2 files changed, 6 insertions(+) diff --git a/Audio-Transcription-Chrome/options.js b/Audio-Transcription-Chrome/options.js index 316d5b6..9f685b7 100644 --- a/Audio-Transcription-Chrome/options.js +++ b/Audio-Transcription-Chrome/options.js @@ -82,6 +82,9 @@ async function startRecord(option) { const socket = new WebSocket(`ws://${option.host}:${option.port}/`); let isServerReady = false; let language = option.language; + if (language === null && !option.multilingual) { + language = 'en'; + } socket.onopen = function(e) { socket.send( JSON.stringify({ diff --git a/Audio-Transcription-Firefox/content.js b/Audio-Transcription-Firefox/content.js index 5a6391e..d849b0d 100644 --- a/Audio-Transcription-Firefox/content.js +++ b/Audio-Transcription-Firefox/content.js @@ -42,6 +42,9 @@ function resampleTo16kHZ(audioData, origSampleRate = 44100) { function startRecording(data) { socket = new WebSocket(`ws://${data.host}:${data.port}/`); language = data.language; + if (language === null && !data.useMultilingual) { + language = 'en'; + } socket.onopen = function(e) { socket.send( JSON.stringify({