Update app.py

app.py

@@ -76,7 +76,7 @@ current_model_path = current_model.path
 if is_colab:
   pipe = StableDiffusionPipeline.from_pretrained(
       current_model.path,
-      torch_dtype=torch.float16,
+      torch_dtype=torch.get_default_dtype(),
       scheduler=DPMSolverMultistepScheduler.from_pretrained(current_model.path, subfolder="scheduler"),
       safety_checker=lambda images, clip_input: (images, False)
       )
@@ -84,13 +84,20 @@ if is_colab:
 else:
   pipe = StableDiffusionPipeline.from_pretrained(
       current_model.path,
-      torch_dtype=torch.float16,
+      torch_dtype=torch.get_default_dtype(),
       scheduler=DPMSolverMultistepScheduler.from_pretrained(current_model.path, subfolder="scheduler")
       )
     
-if torch.cuda.is_available():
-  pipe = pipe.to("cuda")
-  pipe.enable_xformers_memory_efficient_attention()
+to_cuda(torch, pipe)
+
+def to_cuda(torch, pipe):
+    try:
+        if torch.cuda.is_available():
+          pipe = pipe.to("cuda")
+          pipe.enable_xformers_memory_efficient_attention()
+        return True
+    except:
+        return False
 
 device = "GPU 🔥" if torch.cuda.is_available() else "CPU 🥶"
 
@@ -165,22 +172,29 @@ def txt_to_img(model_path, prompt, n_images, neg_prompt, guidance, steps, width,
         if is_colab or current_model == custom_model:
           pipe = StableDiffusionPipeline.from_pretrained(
               current_model_path,
-              torch_dtype=torch.float16,
+              torch_dtype=torch.get_default_dtype(),
               scheduler=DPMSolverMultistepScheduler.from_pretrained(current_model.path, subfolder="scheduler"),
               safety_checker=lambda images, clip_input: (images, False)
               )
         else:
           pipe = StableDiffusionPipeline.from_pretrained(
               current_model_path,
-              torch_dtype=torch.float16,
+              torch_dtype=torch.get_default_dtype(),
               scheduler=DPMSolverMultistepScheduler.from_pretrained(current_model.path, subfolder="scheduler")
               )
           # pipe = pipe.to("cpu")
           # pipe = current_model.pipe_t2i
 
+to_cuda(torch, pipe)
+
+def to_cuda(torch, pipe):
+    try:
         if torch.cuda.is_available():
           pipe = pipe.to("cuda")
           pipe.enable_xformers_memory_efficient_attention()
+        return True
+    except:
+        return False
         last_mode = "txt2img"
 
     prompt = current_model.prefix + prompt  
@@ -214,22 +228,29 @@ def img_to_img(model_path, prompt, n_images, neg_prompt, img, strength, guidance
         if is_colab or current_model == custom_model:
           pipe = StableDiffusionImg2ImgPipeline.from_pretrained(
               current_model_path,
-              torch_dtype=torch.float16,
+              torch_dtype=torch.get_default_dtype(),
               scheduler=DPMSolverMultistepScheduler.from_pretrained(current_model.path, subfolder="scheduler"),
               safety_checker=lambda images, clip_input: (images, False)
               )
         else:
           pipe = StableDiffusionImg2ImgPipeline.from_pretrained(
               current_model_path,
-              torch_dtype=torch.float16,
+              torch_dtype=torch.get_default_dtype(),
               scheduler=DPMSolverMultistepScheduler.from_pretrained(current_model.path, subfolder="scheduler")
               )
           # pipe = pipe.to("cpu")
           # pipe = current_model.pipe_i2i
         
+to_cuda(torch, pipe)
+
+def to_cuda(torch, pipe):
+    try:
         if torch.cuda.is_available():
           pipe = pipe.to("cuda")
           pipe.enable_xformers_memory_efficient_attention()
+        return True
+    except:
+        return False
         last_mode = "img2img"
 
     prompt = current_model.prefix + prompt