Create app. py

app. py

@@ -0,0 +1,64 @@
+import gradio as gr
+import os
+import torch
+import whisper
+from transformers import pipeline
+import subprocess
+
+# =========================
+# 1. LOAD MODELS
+# =========================
+asr_model = whisper.load_model("base")
+
+translator = pipeline(
+    "translation",
+    model="facebook/nllb-200-distilled-600M"
+)
+
+tts = pipeline(
+    "text-to-speech",
+    model="microsoft/speecht5_tts"
+)
+
+# =========================
+# 2. MAIN FUNCTION
+# =========================
+def translate_song(audio_file):
+
+    # STEP 1: TRANSCRIBE (vocals → text)
+    result = asr_model.transcribe(audio_file)
+    kinyarwanda_text = result["text"]
+
+    # STEP 2: TRANSLATE
+    translated = translator(
+        kinyarwanda_text,
+        src_lang="kin_Latn",
+        tgt_lang="eng_Latn"
+    )[0]["translation_text"]
+
+    # STEP 3: TEXT TO SPEECH (new voice)
+    speech = tts(translated)
+
+    output_audio = "translated.wav"
+    with open(output_audio, "wb") as f:
+        f.write(speech["audio"])
+
+    return output_audio, kinyarwanda_text, translated
+
+
+# =========================
+# 3. GRADIO UI
+# =========================
+app = gr.Interface(
+    fn=translate_song,
+    inputs=gr.Audio(type="filepath"),
+    outputs=[
+        gr.Audio(),
+        gr.Textbox(label="Original Lyrics"),
+        gr.Textbox(label="Translated Lyrics")
+    ],
+    title="🎵 AI Song Translator (Kinyarwanda → English)",
+    description="Upload a song in Kinyarwanda and get translated version as audio"
+)
+
+app.launch()