Update app.py

app.py

@@ -1,6 +1,6 @@
-from transformers import pipeline, TFAutoModelForCausalLM, AutoTokenizer
-import soundfile as sf
-import torch
+from transformers import pipeline
+
+tts = pipeline("text-to-speech", model="facebook/fastspeech2-en-ljspeech")
 
 # Initialize the translation pipeline for Russian to English
 translator = pipeline("translation_ru_to_en", model="Helsinki-NLP/opus-mt-ru-en")
@@ -20,14 +20,9 @@ summary = summarizer(translation, max_length=140, min_length=110, do_sample=Fals
 
 print("Summary: ", summary)
 
-tokenizer = AutoTokenizer.from_pretrained("facebook/fastspeech2-en-ljspeech")
-model = TFAutoModelForCausalLM.from_pretrained("facebook/fastspeech2-en-ljspeech")
-
-inputs = tokenizer(summary, return_tensors="tf")
-
-# Generate speech
-with torch.no_grad():
-    logits = model.generate(**inputs)
+speech = tts(summary)
 
-# Save the audio
-sf.write('output_audio.wav', logits.numpy(), samplerate=16000)
+# The output is a list of PyTorch tensors containing the audio data
+# Let's save the first (and only) audio sample to a file
+with open("output1.wav", "wb") as f:
+    f.write(speech[0]["file"].read())