up

run_demo.py

@@ -7,8 +7,6 @@ import librosa
 from transformers import pipeline
 from transformers.utils.logging import disable_progress_bar
 
-SAMPLE_RATE = 16_000
-
 warnings.filterwarnings("ignore")
 
 disable_progress_bar()
@@ -20,7 +18,10 @@ logging.basicConfig(
 logger = logging.getLogger(__name__)
 logger.setLevel(logging.DEBUG)
 
-pipe = pipeline(model="bofenghuang/asr-wav2vec2-french")
+MODEL_NAME = "bofenghuang/asr-wav2vec2-ctc-french"
+SAMPLE_RATE = 16_000
+
+pipe = pipeline(model=MODEL_NAME)
 logger.info("ASR pipeline has been initialized")
 
 
@@ -74,7 +75,7 @@ iface = gr.Interface(
     layout="horizontal",
     # theme="huggingface",
     title="Speech-to-Text in French",
-    description="Realtime demo for French automatic speech recognition.",
+    description=f"Realtime demo for French automatic speech recognition. Demo uses the the fine-tuned checkpoint [{MODEL_NAME}](https://huggingface.co/{MODEL_NAME}) and 🤗 Transformers to transcribe audio files of arbitrary length.",
     allow_flagging="never",
 )
 

run_demo_file.py

@@ -16,7 +16,7 @@ logging.basicConfig(
 logger = logging.getLogger(__name__)
 logger.setLevel(logging.DEBUG)
 
-pipe = pipeline(model="bofenghuang/asr-wav2vec2-french")
+pipe = pipeline(model="bofenghuang/asr-wav2vec2-ctc-french")
 logger.info("ASR pipeline has been initialized")
 
 

run_demo_microphone.py

@@ -16,7 +16,7 @@ logging.basicConfig(
 logger = logging.getLogger(__name__)
 logger.setLevel(logging.DEBUG)
 
-pipe = pipeline(model="bofenghuang/asr-wav2vec2-french")
+pipe = pipeline(model="bofenghuang/asr-wav2vec2-ctc-french")
 logger.info("ASR pipeline has been initialized")
 
 

run_demo_microphone_streaming.py

@@ -16,7 +16,7 @@ logging.basicConfig(
 logger = logging.getLogger(__name__)
 logger.setLevel(logging.DEBUG)
 
-pipe = pipeline(model="bofenghuang/asr-wav2vec2-french")
+pipe = pipeline(model="bofenghuang/asr-wav2vec2-ctc-french")
 logger.info("ASR pipeline has been initialized")