Update app.py

app.py

@@ -5,6 +5,11 @@ import paddlehub as hub
 import urllib
 import cv2
 
+import torch
+
+from spectro import wav_bytes_from_spectrogram_image
+from diffusers import StableDiffusionPipeline
+
 import io
 from os import path
 from pydub import AudioSegment
@@ -13,13 +18,8 @@ from moviepy.editor import *
 import mutagen
 from mutagen.mp3 import MP3
 
-#text_to_img = gr.Blocks.load(name="spaces/DGSpitzer/ERNIE-ViLG", api_key = os.environ.get("hf"))
-text_to_img = gr.Blocks.load(name="spaces/PaddlePaddle/ERNIE-ViLG")
 img_to_text = gr.Blocks.load(name="spaces/pharma/CLIP-Interrogator")
-text_to_music = gr.Interface.load("spaces/fffiloni/text-2-music") 
 
-model = hub.Module(name='ernie_vilg', ak = os.environ.get("ak"), sk = os.environ.get("sk"))
-#model = hub.Module(name='ernie_vilg')
 language_translation_model = hub.Module(name='baidu_translate')
 language_recognition_model = hub.Module(name='baidu_language_recognition')
 
@@ -32,6 +32,11 @@ tips = {"en": "Tips: The input text will be translated into Chinese for generati
 
 count = 0
 
+model_id2 = "riffusion/riffusion-model-v1"
+pipe2 = StableDiffusionPipeline.from_pretrained(model_id2, torch_dtype=torch.float16)
+pipe2 = pipe2.to("cuda")
+
+
 def translate_language_example(text_prompts, style_indx):
     return translate_language(text_prompts)