Upload 35 files

.gitattributes

@@ -33,3 +33,9 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+Samples/emp2.wav filter=lfs diff=lfs merge=lfs -text
+Samples/Montreal[[:space:]]Vacation[[:space:]]Travel[[:space:]]Guide[[:space:]][[:space:]]Expedia.mp3 filter=lfs diff=lfs merge=lfs -text
+Samples/Que[[:space:]]es[[:space:]]TED[[:space:]]y[[:space:]]TEDx.mp3 filter=lfs diff=lfs merge=lfs -text
+Samples/test_mixture.wav filter=lfs diff=lfs merge=lfs -text
+Samples/Gujarati_2.wav filter=lfs diff=lfs merge=lfs -text
+Samples/Telugu_2.wav filter=lfs diff=lfs merge=lfs -text

README.md

@@ -1,10 +1,10 @@
 ---
 title: Pywhisper CPP ASR
-emoji: 📊
-colorFrom: blue
-colorTo: indigo
+emoji: 🦀
+colorFrom: green
+colorTo: green
 sdk: gradio
-sdk_version: 4.1.1
+sdk_version: 4.0.2
 app_file: app.py
 pinned: false
 license: mit

Samples/Gujarati_2.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3d087ac4bc15ddccecbbded4c8fdd2c9501bd8e08c6550f9c03ec46bd2df64da
+size 1597484

Samples/Montreal Vacation Travel Guide Expedia.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bc5148bfd0aca841584c6120c9b02c4c3d9aada2e9b015827a7869803b6471cb
+size 8058808

Samples/Que es TED y TEDx.mp3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:120a8f591bec8715dea618a14bf0538ec03444aaf875994bac0d8a1506f2ffdc
+size 1852114

Samples/Telugu_2.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ccee4c1d338d4d6fc96dfca5f48ecac3d9b9517e44f0302037fe46304e9e76b0
+size 1122348

Samples/emp2.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:133ac69ac8a7d99f1fe92b7dbd810d7f4e44cd4f282179aa8c5d83183e64ad61
+size 1122066

Samples/test_mixture.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7126269f058724336d28bf6e82fa14b7030de321dbafa48c9fa20884d35dab9d
+size 1546250

app.py

@@ -0,0 +1,54 @@
+import gradio as gr
+import soundfile as sf
+import tempfile
+import shutil
+import os
+import librosa
+import time
+import numpy as np
+import subprocess 
+from pywhispercpp.model import Model
+
+# model = Model('base.en', n_threads=6,models_dir="./Models") # Only english
+# model = Model('base', n_threads=6,models_dir="./Models",language="hindi",translate=False)  # Multilingual
+model = Model('medium', n_threads=6,models_dir="./Models",language="hindi",translate=False)  # Multilingual
+
+def resample_to_16k(audio, orig_sr):
+    y_resampled = librosa.resample(y=audio, orig_sr=orig_sr, target_sr = 16000)
+    return y_resampled
+
+def transcribe(audio):
+    print(type(audio))
+    sr,y = audio
+    y = y.astype(np.float32)
+    y /= np.max(np.abs(y))
+    y_resampled = resample_to_16k(y, sr)
+    
+    
+    with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as temp_audio:
+        temp_audio_path = temp_audio.name
+        sf.write(temp_audio_path, y_resampled, 16000)
+
+    start_time_py = time.time()
+    py_result = model.transcribe(f'{temp_audio_path}', n_threads=6)
+    end_time_py = time.time()
+    print("Py_result : ",py_result)
+    print("--------------------------")
+    print(f"Execution time using py: {end_time_py - start_time_py} seconds")
+    output_text = ""
+    for segment in py_result:
+        output_text+=segment.text
+    return output_text, (end_time_py - start_time_py)
+
+
+
+demo = gr.Interface(
+    transcribe,
+    inputs = "microphone",
+    # gr.Audio(sources=["microphone"]),
+    outputs=[gr.Textbox(label="Py_Transcription"),gr.Textbox(label="Time taken for Transcription")],
+    # examples=["./Samples/Hindi_1.mp3","./Samples/Hindi_2.mp3","./Samples/Tamil_1.mp3","./Samples/Tamil_2.mp3","./Samples/Marathi_1.mp3","./Samples/Marathi_2.mp3","./Samples/Nepal_1.mp3","./Samples/Nepal_2.mp3","./Samples/Telugu_1.wav","./Samples/Telugu_2.wav","./Samples/Malayalam_1.wav","./Samples/Malayalam_2.wav","./Samples/Gujarati_1.wav","./Samples/Gujarati_2.wav","./Samples/Bengali_1.wav","./Samples/Bengali_2.wav"]
+    examples=["./Samples/Hindi_1.mp3","./Samples/Hindi_2.mp3","./Samples/Hindi_3.mp3","./Samples/Hindi_4.mp3","./Samples/Hindi_5.mp3"] # only hindi   # ,"./Samples/Tamil_1.mp3","./Samples/Tamil_2.mp3","./Samples/Marathi_1.mp3","./Samples/Marathi_2.mp3","./Samples/Nepal_1.mp3","./Samples/Nepal_2.mp3","./Samples/Telugu_1.wav","./Samples/Telugu_2.wav","./Samples/Malayalam_1.wav","./Samples/Malayalam_2.wav","./Samples/Gujarati_1.wav","./Samples/Gujarati_2.wav","./Samples/Bengali_1.wav","./Samples/Bengali_2.wav"]
+)
+
+demo.launch()

requirements.txt

@@ -0,0 +1,3 @@
+librosa
+soundfile
+pywhispercpp