Update app.py

app.py

@@ -275,7 +275,7 @@ def captioning(img):
     '''
     generated_ids = model5.generate(
         **inputsa,
-        do_sample=text_decoding_method == "Nucleus sampling",
+        text_decoding_method == "Nucleus sampling",
         num_beams=1,
         max_length=128,
         min_length=64,
@@ -284,7 +284,6 @@ def captioning(img):
         length_penalty=2.0,
         temperature=0.5,
     )
-            
 
     generated_text = processor5.batch_decode(generated_ids, skip_special_tokens=True)[0].strip()
     generated_text = generated_text.replace(cap_prompt, "").strip() #Or could try .split(prompt, 1)[-1].strip()
@@ -296,7 +295,7 @@ def captioning(img):
         #with torch.no_grad():
         generated_ids = model5.generate(
             **inputs,
-            do_sample=text_decoding_method == "Nucleus sampling",
+            text_decoding_method == "Nucleus sampling",
             num_beams=1,
             max_length=64,
             #min_length=16,
@@ -316,7 +315,7 @@ def captioning(img):
     ).to('cuda')
     generated_ids = model5.generate(
             **inputf,
-            do_sample=text_decoding_method == "Nucleus sampling",
+            text_decoding_method == "Nucleus sampling",
             num_beams=1,
             max_length=96,
             min_length=64,
@@ -427,35 +426,35 @@ def generate_30(
     generator = torch.Generator(device='cuda').manual_seed(seed)
     if latent_file is not None:  # Check if a latent file is provided
         sd_image_a = Image.open(latent_file.name).convert('RGB')
-        sd_image_a.resize((224,224), Image.LANCZOS)
+        sd_image_a.resize((768,768), Image.LANCZOS)
         #sd_image_a.resize((height,width), Image.LANCZOS)
         caption=[]
         caption.append(captioning(sd_image_a))
         if latent_file_2 is not None:  # Check if a latent file is provided
             sd_image_b = Image.open(latent_file_2.name).convert('RGB')
             #sd_image_b.resize((height,width), Image.LANCZOS)
-            sd_image_b.resize((224,224), Image.LANCZOS)
+            sd_image_b.resize((768,768), Image.LANCZOS)
             caption.append(captioning(sd_image_b))
         else:
             sd_image_b = None
         if latent_file_3 is not None:  # Check if a latent file is provided
             sd_image_c = Image.open(latent_file_3.name).convert('RGB')
             #sd_image_c.resize((height,width), Image.LANCZOS)
-            sd_image_c.resize((224,224), Image.LANCZOS)
+            sd_image_c.resize((768,768), Image.LANCZOS)
             caption.append(captioning(sd_image_c))
         else:
             sd_image_c = None
         if latent_file_4 is not None:  # Check if a latent file is provided
             sd_image_d = Image.open(latent_file_4.name).convert('RGB')
             #sd_image_d.resize((height,width), Image.LANCZOS)
-            sd_image_d.resize((224,224), Image.LANCZOS)
+            sd_image_d.resize((768,768), Image.LANCZOS)
             caption.append(captioning(sd_image_d))
         else:
             sd_image_d = None
         if latent_file_5 is not None:  # Check if a latent file is provided
             sd_image_e = Image.open(latent_file_5.name).convert('RGB')
             #sd_image_e.resize((height,width), Image.LANCZOS)
-            sd_image_e.resize((224,224), Image.LANCZOS)
+            sd_image_e.resize((768,768), Image.LANCZOS)
             caption.append(captioning(sd_image_e))
         else:
             sd_image_e = None