Update app.py

app.py

@@ -20,7 +20,7 @@ from typing import Tuple
 import paramiko
 import datetime
 # import cyper
-from image_gen_aux import UpscaleWithModel
+#from image_gen_aux import UpscaleWithModel
 from huggingface_hub import hf_hub_download
 
 torch.backends.cuda.matmul.allow_tf32 = False
@@ -85,7 +85,7 @@ neg_prompt_2 = " 'non-photorealistic':1.5, 'unrealistic skin','unattractive face
 
 device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
 
-upscaler = UpscaleWithModel.from_pretrained("Kim2091/ClearRealityV1").to(device)
+#upscaler = UpscaleWithModel.from_pretrained("Kim2091/ClearRealityV1").to(device)
 
 def load_and_prepare_model():
     vaeX=AutoencoderKL.from_pretrained("ford442/stable-diffusion-3.5-large-fp32", safety_checker=None, use_safetensors=True, subfolder='vae', low_cpu_mem_usage=False, torch_dtype=torch.float32, token=True)
@@ -233,17 +233,18 @@ def generate(
         #image.save(output_image_file,optimize=False,compress_level=0)
         #upload_to_ftp(output_image_file) 
         #upscaler.to(torch.device('cuda'))
-        with torch.no_grad():
-            upscale2 = upscaler(image, tiling=True, tile_width=256, tile_height=256)
+        #with torch.no_grad():
+        #    upscale2 = upscaler(image, tiling=True, tile_width=256, tile_height=256)
         #timestamp = datetime.datetime.now().strftime("%Y%m%d_%H%M%S")
-        print('-- got upscaled image --')
-        downscale2 = upscale2.resize((upscale2.width // 4, upscale2.height // 4),Image.LANCZOS)
-        upscale_path = f"Large_Lora_L2_SR{seed}.png"
-        downscale2.save(upscale_path,optimize=False,compress_level=0)
-        upload_to_ftp(upscale_path) 
+        #print('-- got upscaled image --')
+        #downscale2 = upscale2.resize((upscale2.width // 4, upscale2.height // 4),Image.LANCZOS)
+        #upscale_path = f"Large_Lora_L2_SR{seed}.png"
+        #downscale2.save(upscale_path,optimize=False,compress_level=0)
+        image.save(output_image_file,optimize=False,compress_level=0)
+        upload_to_ftp(output_image_file) 
         #uploadNote(prompt,num_inference_steps,guidance_scale,timestamp)
         unique_name = str(uuid.uuid4()) + ".png"  
-        os.symlink(upscale_path, unique_name)  
+        os.symlink(output_image_file, unique_name)  
         return [unique_name], seed
     else:
         original_prompt_embeds_cpu = prompt_embeds.cpu()