Update myapp.py

myapp.py

@@ -1,56 +1,56 @@
-from flask import Flask, request, send_file
+from flask import Flask, jsonify, request, send_file
 from flask_cors import CORS
 import torch
 from diffusers import DiffusionPipeline
-import io
 import numpy as np
 import random
+import io
+from PIL import Image
 
 # Initialize the Flask app
 myapp = Flask(__name__)
 CORS(myapp)  # Enable CORS if needed
 
 # Load the model
-device = "cuda" if torch.cuda.is_available() else "cpu"
+device = "cpu"
+dtype = torch.float16
 
-# Load the DiffusionPipeline for "prompthero/openjourney-v4"
-pipe = DiffusionPipeline.from_pretrained("prompthero/openjourney-v4").to(device)
+repo = "prompthero/openjourney-v4"
+pipe = DiffusionPipeline.from_pretrained(repo, torch_dtype=dtype).to(device)
 
-# Define max values for seed and image size
 MAX_SEED = np.iinfo(np.int32).max
 MAX_IMAGE_SIZE = 1344
 
-@app.route('/')
+@myapp.route('/')  # Use 'myapp' instead of 'app'
 def home():
-    return "Welcome to the OpenJourney Image Generation API!"
+    return "Welcome to the Diffusion Image Generation API!"  # Basic home response
 
-@app.route('/generate_image', methods=['POST'])
+@myapp.route('/generate_image', methods=['POST'])
 def generate_image():
     data = request.json
     
     # Get inputs from request JSON
-    prompt = data.get('prompt', 'Astronaut in a jungle, cold color palette, muted colors, detailed, 8k')
+    prompt = data.get('prompt', '')
     negative_prompt = data.get('negative_prompt', None)
     seed = data.get('seed', 0)
     randomize_seed = data.get('randomize_seed', True)
     width = data.get('width', 1024)
     height = data.get('height', 1024)
-    guidance_scale = data.get('guidance_scale', 7.5)  # Default to a higher guidance scale for better results
-    num_inference_steps = data.get('num_inference_steps', 50)  # Default number of steps
-
+    guidance_scale = data.get('guidance_scale', 5.0)
+    num_inference_steps = data.get('num_inference_steps', 28)
+    
     # Randomize seed if requested
     if randomize_seed:
         seed = random.randint(0, MAX_SEED)
-
+    
     # Generate the image
     generator = torch.Generator().manual_seed(seed)
     image = pipe(
         prompt=prompt,
-        negative_prompt=negative_prompt,
-        width=width,
-        height=height,
         guidance_scale=guidance_scale,
         num_inference_steps=num_inference_steps,
+        width=width,
+        height=height,
         generator=generator
     ).images[0]