Delete README.Md

README.Md

@@ -1,23 +0,0 @@
-## Usage
-The model can be used directly (without a language model) as follows:
-```python
-import soundfile as sf
-import torch
-from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
-import argparse
-def parse_transcription(wav_file):
-    # load pretrained model
-    processor = Wav2Vec2Processor.from_pretrained("Harveenchadha/vakyansh-wav2vec2-hindi-him-4200")
-    model = Wav2Vec2ForCTC.from_pretrained("Harveenchadha/vakyansh-wav2vec2-hindi-him-4200")
-    # load audio
-    audio_input, sample_rate = sf.read(wav_file)
-    # pad input values and return pt tensor
-    input_values = processor(audio_input, sampling_rate=sample_rate, return_tensors="pt").input_values
-    # INFERENCE
-    # retrieve logits & take argmax
-    logits = model(input_values).logits
-    predicted_ids = torch.argmax(logits, dim=-1)
-    # transcribe
-    transcription = processor.decode(predicted_ids[0], skip_special_tokens=True)
-    print(transcription)
-```