hrishikesh commited on
Commit
9f9fc69
1 Parent(s): a26c9c1

Create app.py

Browse files
Files changed (1) hide show
  1. app.py +76 -0
app.py ADDED
@@ -0,0 +1,76 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # -*- coding: utf-8 -*-
2
+ """Image & Music Generator.ipynb
3
+
4
+ Automatically generated by Colaboratory.
5
+
6
+ Original file is located at
7
+ https://colab.research.google.com/drive/1TBH0E9wF716_A8o_yT68mSp_IFnb14_F
8
+ """
9
+
10
+ !nvidia-smi
11
+
12
+ pip install accelerate
13
+
14
+ !pip install -q https://github.com/camenduru/stable-diffusion-webui-colab/releases/download/0.0.15/xformers-0.0.15.dev0+189828c.d20221207-cp38-cp38-linu
15
+
16
+ ! pip install -U transformers diffusers gradio ftfy pydub -q
17
+
18
+ !wget https://raw.githubusercontent.com/hmartiro/riffusion-inference/main/riffusion/audio.py
19
+
20
+ import gradio as gr
21
+
22
+ import torch
23
+ from diffusers import StableDiffusionPipeline
24
+ from audio import wav_bytes_from_spectrogram_image
25
+
26
+ model_id = "riffusion/riffusion-model-v1"
27
+ pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16)
28
+
29
+ pipe = pipe.to("cuda")
30
+ #pipe.enable_xformers_memory_efficient_attention()
31
+
32
+ import random
33
+ COLORS = [
34
+ ["#ff0000", "#00ff00"],
35
+ ["#00ff00", "#0000ff"],
36
+ ["#0000ff", "#ff0000"],
37
+ ]
38
+
39
+ from diffusers import StableDiffusionPipeline
40
+ import torch
41
+
42
+ img_model_id = "runwayml/stable-diffusion-v1-5"
43
+ img_pipe = StableDiffusionPipeline.from_pretrained(img_model_id, torch_dtype=torch.float16, revision="fp16")
44
+ img_pipe = img_pipe.to("cuda")
45
+ #img_pipe.enable_xformers_memory_efficient_attention()
46
+
47
+ prompt = 'morning sunshine'
48
+
49
+ spectogram = pipe(prompt).images[0]
50
+ wav = wav_bytes_from_spectrogram_image(spectogram)
51
+ with open("output.wav", "wb") as f:
52
+ f.write(wav[0].getbuffer())
53
+
54
+ def audio_gen(prompt):
55
+ spectogram = pipe(prompt).images[0]
56
+ wav = wav_bytes_from_spectrogram_image(spectogram)
57
+ with open("output.wav", "wb") as f:
58
+ f.write(wav[0].getbuffer())
59
+ print("audio saved")
60
+
61
+ return ('output.wav',)
62
+
63
+ audio_gen("lazy nights")
64
+
65
+ gr.Interface(
66
+ audio_gen,
67
+ inputs=[gr.Textbox(label="prompt")],
68
+ outputs=[
69
+ gr.Audio(type='filepath')
70
+
71
+ ],
72
+ title = 'Music Generator'
73
+ ).launch(debug = True)
74
+
75
+
76
+