app commit with wav2vec-base-960h

app.py

@@ -1,7 +1,44 @@
+import soundfile as sf
+import torch
+from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
+from transformers import M2M100ForConditionalGeneration, M2M100Tokenizer
 import gradio as gr
+import sox
 
-def greet(name):
-    return "Hello " + name + "!!"
+def convert(inputfile, outfile):
+    sox_tfm = sox.Transformer()
+    sox_tfm.set_output_format(
+        file_type="wav", channels=1, encoding="signed-integer", rate=16000, bits=16
+    )
+    sox_tfm.build(inputfile, outfile)
+    
+    
+    
+processor = Wav2Vec2Processor.from_pretrained("facebook/wav2vec2-base-960h")
+model = Wav2Vec2ForCTC.from_pretrained("facebook/wav2vec2-base-960h")
 
-iface = gr.Interface(fn=greet, inputs="text", outputs="text")
-iface.launch()
+def parse_transcription(wav_file):
+    filename = wav_file.name.split('.')[0]
+    convert(wav_file.name, filename + "16k.wav")
+    speech, _ = sf.read(filename + "16k.wav")
+    input_values = processor(speech, sampling_rate=16_000, return_tensors="pt").input_values
+    logits = model(input_values).logits
+    predicted_ids = torch.argmax(logits, dim=-1)
+    transcription = processor.decode(predicted_ids[0], skip_special_tokens=True)
+    return transcription,
+
+
+
+output1 = gr.outputs.Textbox(label="Transcription in English: ")
+output2 = gr.outputs.Textbox(label="Validated Transcription in English")
+    
+input_ = gr.inputs.Audio(source="microphone", type="file") 
+#gr.Interface(parse_transcription, inputs = input_,  outputs="text", 
+#             analytics_enabled=False, show_tips=False, enable_queue=True).launch(inline=False);
+             
+gr.Interface(parse_transcription, inputs = input_,  outputs=[output1, output2], analytics_enabled=False, 
+                                                                            show_tips=False, 
+                                                                            theme='huggingface',
+                                                                            layout='vertical',
+                                                                            title="Piecurus Test on Speech Transcription",
+                                                                            description="This is a live demo for Speech to Text Translation. Models used: facebook/wav2vec2-base-960h", enable_queue=True).launch( inline=False)