adding quantization

app.py

@@ -10,6 +10,10 @@ model_name = "OthmaneJ/distil-wav2vec2"
 processor = Wav2Vec2Processor.from_pretrained(model_name)
 model = Wav2Vec2ForCTC.from_pretrained(model_name)
 
+# quantization
+model.eval()
+model_int8 = torch.quantization.quantize_dynamic(model, dtype=torch.qint8,inplace = True,)
+
 # define function to read in sound file
 # def map_to_array(file):
 #    speech, sample_rate = sf.read(file)
@@ -29,7 +33,7 @@ def inference(audio):
 
 inputs = gr.inputs.Audio(label="Input Audio", type="file")
 outputs =  gr.outputs.Textbox(label="Output Text")
-title = "Distilled wav2vec 2.0"
+title = "distilled wav2vec 2.0 (with quantization)"
 description = "Gradio demo for Robust wav2vec 2.0. To use it, simply upload your audio, or click one of the examples to load them. Read more at the links below. Currently supports .wav and .flac files"
 article = "<p style='text-align: center'><a href='https://arxiv.org/abs/2104.01027' target='_blank'>Robust wav2vec 2.0: Analyzing Domain Shift in Self-Supervised Pre-Training</a> | <a href='https://github.com/pytorch/fairseq' target='_blank'>Github Repo</a></p>"
 examples=[['poem.wav']]