Update app.py

app.py

@@ -1,6 +1,4 @@
 import os
-import time
-import requests
 from huggingface_hub import login
 import torch
 import torchaudio
@@ -9,20 +7,20 @@ import gradio as gr
 from stable_audio_tools import get_pretrained_model
 from stable_audio_tools.inference.generation import generate_diffusion_cond
 
-# Authenticate Hugging Face Hub
+# Authenticate
 token = os.getenv("HUGGINGFACE_TOKEN")
 if not token:
     raise RuntimeError("HUGGINGFACE_TOKEN not set")
 login(token=token, add_to_git_credential=False)
 
-# Load audio model
+# Load model
 device = "cuda" if torch.cuda.is_available() else "cpu"
 model, config = get_pretrained_model("stabilityai/stable-audio-open-small")
 model = model.to(device)
 sample_rate = config["sample_rate"]
 sample_size = config["sample_size"]
 
-# Audio generation function
+# Inference function
 def generate_audio(prompt):
     conditioning = [{"prompt": prompt, "seconds_total": 11}]
     with torch.no_grad():
@@ -39,70 +37,22 @@ def generate_audio(prompt):
     torchaudio.save(path, output, sample_rate)
     return path
 
-# Image generation function using Replicate
-def generate_image(prompt):
-    replicate_token = os.getenv("REPLICATE_API_TOKEN")
-    if not replicate_token:
-        raise RuntimeError("REPLICATE_API_TOKEN not set")
-
-    url = "https://api.replicate.com/v1/predictions"
-    headers = {
-        "Authorization": f"Token {replicate_token}",
-        "Content-Type": "application/json"
-    }
-    data = {
-        "version": "5ee6b41748a4e3e3d3a212ed4a29379d6a13b9265fd00fe59e28c2767a5e82eb",
-        "input": {
-            "prompt": prompt,
-            "style": "surreal"
-        }
-    }
-    response = requests.post(url, headers=headers, json=data)
-    response.raise_for_status()
-    prediction = response.json()
-
-    status = prediction["status"]
-    get_url = prediction["urls"]["get"]
-
-    while status not in ["succeeded", "failed"]:
-        time.sleep(1.5)
-        resp = requests.get(get_url, headers=headers)
-        prediction = resp.json()
-        status = prediction["status"]
-    
-    if status != "succeeded":
-        raise RuntimeError(f"Image generation failed: {prediction}")
-    
-    image_url = prediction["output"]
-    image_path = "output.png"
-    image_data = requests.get(image_url).content
-    with open(image_path, "wb") as f:
-        f.write(image_data)
-    
-    return image_path
-
-# Combined generation function
-def generate_assets(prompt):
-    audio_path = generate_audio(prompt)
-    image_path = generate_image(prompt)
-    return audio_path, image_path
-
-# Gradio UI
+# 🌀 Hot Prompt Club UI
 gr.Interface(
-    fn=generate_assets,
+    fn=generate_audio,
     inputs=gr.Textbox(
-        label="🎤 Prompt your sonic + visual art",
+        label="🎤 Prompt your sonic art here",
         placeholder="e.g. 'drunk driving with mario and yung lean'"
     ),
-    outputs=[
-        gr.Audio(type="filepath", label="🧠 Generated Audio"),
-        gr.Image(type="filepath", label="🎨 Generated Image")
-    ],
+    outputs=gr.Audio(
+        type="filepath",
+        label="🧠 Generated Audio"
+    ),
     title='🌐 Hot Prompts in Your Area: "My Husband Is Dead"',
-    description="Enter a fun sound idea — generate audio *and* visual from one prompt.",
+    description="Enter a fun sound idea for music art.",
     examples=[
-        "ghosts peeing",
-        "Tech startup boss villain entrance music",
-        "Dolphin hootin'"
+        "ghosts peeing in a server room",
+        "tech startup boss villain entrance music",
+        "AI doing acid in a technofeudalist dystopia"
     ]
-).launch()
+).launch()