Amir Zait commited on
Commit
077c45d
1 Parent(s): f7c2e78

fixed pipeline

Browse files
Files changed (1) hide show
  1. app.py +2 -3
app.py CHANGED
@@ -14,8 +14,7 @@ api_token = os.getenv("API_TOKEN")
14
  asr_processor = AutoProcessor.from_pretrained("imvladikon/wav2vec2-xls-r-300m-hebrew")
15
  asr_model = AutoModelForCTC.from_pretrained("imvladikon/wav2vec2-xls-r-300m-hebrew")
16
 
17
- en_he_translator = pipeline("translation_en_to_he")
18
-
19
 
20
  def process_audio_file(file):
21
  data, sr = librosa.load(file)
@@ -59,7 +58,7 @@ def parse_transcription(wav_file):
59
  logits = trans_model(input_values).logits
60
  predicted_ids = torch.argmax(logits, dim=-1)
61
  transcription = trans_processor.decode(predicted_ids[0], skip_special_tokens=True)
62
- translated = en_he_translator(trasncription)
63
  return transcription
64
 
65
  output = gr.outputs.Textbox(label="TEXT")
 
14
  asr_processor = AutoProcessor.from_pretrained("imvladikon/wav2vec2-xls-r-300m-hebrew")
15
  asr_model = AutoModelForCTC.from_pretrained("imvladikon/wav2vec2-xls-r-300m-hebrew")
16
 
17
+ he_en_translator = pipeline("translation", model="Helsinki-NLP/opus-mt-tc-big-he-en")
 
18
 
19
  def process_audio_file(file):
20
  data, sr = librosa.load(file)
 
58
  logits = trans_model(input_values).logits
59
  predicted_ids = torch.argmax(logits, dim=-1)
60
  transcription = trans_processor.decode(predicted_ids[0], skip_special_tokens=True)
61
+ translated = he_en_translator(trasncription)
62
  return transcription
63
 
64
  output = gr.outputs.Textbox(label="TEXT")