Update app.py

app.py

@@ -3,14 +3,19 @@ import gradio as gr
 import requests
 from PIL import Image
 from transformers import BlipProcessor, BlipForConditionalGeneration
-import torch
-import soundfile as sf
-from diffusers import StableAudioPipeline
-import os
-from huggingface_hub import login
+from outetts.v0_1.interface import InterfaceHF, InterfaceGGUF
+
+# Initialize the interface with the Hugging Face model
+interface = InterfaceHF("OuteAI/OuteTTS-0.1-350M")
+
+# Or initialize the interface with a GGUF model
+# interface = InterfaceGGUF("path/to/model.gguf")
+
+
+
+
 
 
-login(token=os.environ["HF_TOKEN"])
 
 processor = BlipProcessor.from_pretrained("Salesforce/blip-image-captioning-large")
 model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-large").to("cpu")
@@ -41,25 +46,17 @@ with gr.Blocks(theme=gr.themes.Ocean(primary_hue="pink", neutral_hue="indigo", f
       return processor.decode(out[0], skip_special_tokens=True)
 
     def leer(texto):
-        prompt = texto
-        negative_prompt = "Low quality."
-
-        generator = torch.Generator("cpu").manual_seed(0)
-
-        audio = pipe(
-            prompt,
-            negative_prompt=negative_prompt,
-            num_inference_steps=200,
-            audio_end_in_s=10.0,
-            num_waveforms_per_prompt=3,
-            generator=generator,
-        ).audios
+        output = interface.generate(
+        text= texto,
+        temperature=0.1,
+        repetition_penalty=1.1,
+        max_length=4096
+        )
+        
+        #output.play()
+        #output.save("output.wav")
         
-        salida = audio[0].T.float().cpu().numpy()
-        print("Fijar audio")
-        sf.write("demo.wav", salida, pipe.vae.sampling_rate)
-        print("sf writed")
-        return sf.read("demo.wav")
+        return output
 
 
     button.click(describir, [textbox], output)