Update app.py

app.py

@@ -52,17 +52,17 @@ current_model = models[1] if is_colab else models[0]
 current_model_path = current_model.path
 
 if is_colab:
-  pipe = StableDiffusionPipeline.from_pretrained(current_model.path, torch_dtype=torch.float16, scheduler=scheduler, safety_checker=lambda images, clip_input: (images, False))
+  pipe = StableDiffusionPipeline.from_pretrained(current_model.path, scheduler=scheduler, safety_checker=lambda images, clip_input: (images, False))
 
-else: # download all models
+else: # download all models All models needed torch_dtype=torch.float16 removed for CPU. Add a If/Else so you can use both
   print(f"{datetime.datetime.now()} Downloading vae...")
-  vae = AutoencoderKL.from_pretrained(current_model.path, subfolder="vae", torch_dtype=torch.float16)
+  vae = AutoencoderKL.from_pretrained(current_model.path, subfolder="vae")
   for model in models:
     try:
         print(f"{datetime.datetime.now()} Downloading {model.name} model...")
-        unet = UNet2DConditionModel.from_pretrained(model.path, subfolder="unet", torch_dtype=torch.float16)
-        model.pipe_t2i = StableDiffusionPipeline.from_pretrained(model.path, unet=unet, vae=vae, torch_dtype=torch.float16, scheduler=scheduler)
-        model.pipe_i2i = StableDiffusionImg2ImgPipeline.from_pretrained(model.path, unet=unet, vae=vae, torch_dtype=torch.float16, scheduler=scheduler)
+        unet = UNet2DConditionModel.from_pretrained(model.path, subfolder="unet")
+        model.pipe_t2i = StableDiffusionPipeline.from_pretrained(model.path, unet=unet, vae=vae, scheduler=scheduler)
+        model.pipe_i2i = StableDiffusionImg2ImgPipeline.from_pretrained(model.path, unet=unet, vae=vae, scheduler=scheduler)
     except Exception as e:
         print(f"{datetime.datetime.now()} Failed to load model " + model.name + ": " + str(e))
         models.remove(model)
@@ -98,7 +98,7 @@ def inference(model_name, prompt, guidance, steps, width=512, height=512, seed=0
       current_model = model
       model_path = current_model.path
 
-  generator = torch.Generator('cuda').manual_seed(seed) if seed != 0 else None
+  generator = torch.Generator('cpu').manual_seed(seed) if seed != 0 else None #not using cuda, I can add another if/else if you like
 
   try:
     if img is not None:
@@ -119,7 +119,7 @@ def txt_to_img(model_path, prompt, neg_prompt, guidance, steps, width, height, g
         current_model_path = model_path
 
         if is_colab or current_model == custom_model:
-          pipe = StableDiffusionPipeline.from_pretrained(current_model_path, torch_dtype=torch.float16, scheduler=scheduler, safety_checker=lambda images, clip_input: (images, False))
+          pipe = StableDiffusionPipeline.from_pretrained(current_model_path, scheduler=scheduler, safety_checker=lambda images, clip_input: (images, False))
         else:
           pipe = pipe.to("cpu")
           pipe = current_model.pipe_t2i
@@ -152,7 +152,7 @@ def img_to_img(model_path, prompt, neg_prompt, img, strength, guidance, steps, w
         current_model_path = model_path
 
         if is_colab or current_model == custom_model:
-          pipe = StableDiffusionImg2ImgPipeline.from_pretrained(current_model_path, torch_dtype=torch.float16, scheduler=scheduler, safety_checker=lambda images, clip_input: (images, False))
+          pipe = StableDiffusionImg2ImgPipeline.from_pretrained(current_model_path, scheduler=scheduler, safety_checker=lambda images, clip_input: (images, False))
         else:
           pipe = pipe.to("cpu")
           pipe = current_model.pipe_i2i
@@ -163,7 +163,7 @@ def img_to_img(model_path, prompt, neg_prompt, img, strength, guidance, steps, w
 
     prompt = current_model.prefix + prompt
     ratio = min(height / img.height, width / img.width)
-    img = img.resize((int(img.width * ratio), int(img.height * ratio)), Image.LANCZOS)
+    img = img.resize((int(img.width * ratio), int(img.height * ratio)), Image.Resampling.LANCZOS) #added Resampling to avoid depreciation error 
     result = pipe(
         prompt,
         negative_prompt = neg_prompt,