Update app.py

app.py

@@ -2,4 +2,51 @@ import pandas as pd
 import gradio as gr
 import torch
 import torchaudio
+import warnings
+from cryptography.utils import CryptographyDeprecationWarning
+with warnings.catch_warnings():
+    warnings.filterwarnings('ignore', category=CryptographyDeprecationWarning)
+    import paramiko
+ torch._C._jit_override_can_fuse_on_cpu(False)
+torch._C._jit_override_can_fuse_on_gpu(False)
+torch._C._jit_set_texpr_fuser_enabled(False)
+torch._C._jit_set_nvfuser_enabled(False)
 
+loader = torch.jit.load("audio_loader.pt")
+model = torch.jit.load('QuartzNet_thunderspeech_3.pt')
+
+vocab = model.text_transform.vocab.itos
+vocab[-1] = ''
+
+def convert_probs(probs):
+  ids = probs.argmax(1)[0]
+  s = []
+  if vocab[ids[0]]: s.append(vocab[ids[0]])
+  for i in range(1,len(ids)):
+    if ids[i-1] != ids[i]:
+      new = vocab[ids[i]]
+      if new: s.append(new)
+  #return '.'.join(s)
+  return s
+ 
+  
+def predict(path):
+  audio = loader(path)
+  probs = model(audio, torch.tensor(audio.shape[0] * [audio.shape[-1]], device=audio.device))[0]
+  return convert_probs(probs)
+ 
+ 
+from difflib import SequenceMatcher
+
+def similar(a, b):
+    return SequenceMatcher(None, a, b).ratio()
+    
+def compare (word_choice, path):
+  etalon = df.loc[df['replica'] == word_choice, 'transcription'].values[0]
+  user = predict(path)
+  similar(user, etalon) 
+
+
+word_choice = gr.inputs.Dropdown(list(df['replica'].unique()), label="Choose a word")
+
+gr.Interface(fn=compare, inputs=[gr.inputs.Audio(source='microphone', type='filepath', optional=True), word_choice], outputs= 'text').launch(debug=True)