Add label

app.py

@@ -5,13 +5,13 @@ from diffusers import StableDiffusionPipeline
 import gradio as gr
 import torch
 
-from spectro import wav_bytes_from_spectrogram_image
-
 device = "cuda" if torch.cuda.is_available() else "cpu"
 dtype = torch.float16 if device == "cuda" else torch.float32
 
 model_id = "runwayml/stable-diffusion-v1-5"
-pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=dtype)
+pipe = StableDiffusionPipeline.from_pretrained(model_id,
+                                               torch_dtype=dtype,
+                                               revision="fp16")
 pipe = pipe.to(device)
 
 model_id2 = "riffusion/riffusion-model-v1"
@@ -29,7 +29,6 @@ title = """
           <div style="display: inline-flex; align-items: center; gap: 0.8rem; font-size: 1.75rem;">
             <h1 style="font-weight: 950; margin-bottom: 7px; color: #000; font-weight: bold;">Riffusion and Stable Diffusion</h1>
           </div>
-          <p style="margin-bottom: 10px; font-size: 98%; color: #666;">Text to music player.</p>
         </div>
         """
 def get_bg_image(prompt):
@@ -39,18 +38,18 @@ def get_bg_image(prompt):
     return image_output
 
 def get_music(prompt):
-    spec = pipe2(prompt).images[0]
+    spec = pipe2(prompt, height=512, width=512).images[0]
     print(spec)
     wav = wav_bytes_from_spectrogram_image(spec)
     with open("output.wav", "wb") as f:
-      f.write(wav[0].getbuffer())
-    return 'output.wav'
+        f.write(wav[0].getbuffer())
+    return "output.wav"
 
 def infer(prompt):
     image = get_bg_image(prompt)
     audio = get_music(prompt)
     return (
-        gr.make_waveform(audio, bg_image=image, bars_color=random.choice(COLORS)),
+        gr.make_waveform("output.wav", bg_image=image, bars_color=random.choice(COLORS)),
     )
 
 css = """
@@ -81,12 +80,13 @@ css = """
 with gr.Blocks(css=css) as demo:
     gr.HTML(title)
     with gr.Column(elem_id="col-container"):
-      prompt_input = gr.Textbox(placeholder="a cat diva singing in a New York jazz club",
+      prompt_input = gr.Textbox(placeholder="The Beatles playing for the queen",
                                 elem_id="prompt-in",
-                                show_label=False)
+                                label="Enter your music prompt")
       with gr.Row(elem_id="btn-container"):
         send_btn = gr.Button(value="Send", elem_id="submit-btn")      
-      video_output = gr.Video()
-      send_btn.click(infer, inputs=[prompt_input], outputs=[video_output])
+      send_btn.click(infer,
+                     inputs=[prompt_input],
+                     outputs=[gr.Video()])
 
-demo.queue().launch(debug=True)
+demo.launch().debug(True)