Update app.py

app.py

@@ -160,9 +160,9 @@ pipe = load_and_prepare_model()
 # text models
 #checkpoint = "microsoft/Phi-3.5-mini-instruct"
 checkpoint = "ford442/Phi-3.5-mini-instruct-bf16"
-#captioner = pipeline(model="ydshieh/vit-gpt2-coco-en",device='cuda', task="image-to-text")
+captioner = pipeline(model="ydshieh/vit-gpt2-coco-en",device='cuda', task="image-to-text")
 #captioner_2 = pipeline(model="Salesforce/blip-image-captioning-base",device='cuda', task="image-to-text")
-captioner_3 = pipeline(model="ford442/blip-image-to-text-large-bf16",device='cuda', task="image-to-text")
+#captioner_3 = pipeline(model="ford442/blip-image-to-text-large-bf16",device='cuda', task="image-to-text")
 model5 = Blip2ForConditionalGeneration.from_pretrained("ford442/blip2-image-to-text-bf16").to('cuda')
 processor5 = Blip2Processor.from_pretrained("ford442/blip2-image-to-text-bf16")
 txt_tokenizer = AutoTokenizer.from_pretrained(checkpoint, add_prefix_space=False)
@@ -352,43 +352,43 @@ def generate_30(
         sd_image_a.resize((height,width), Image.LANCZOS)
         caption=[]
         caption_2=[]
-        #caption.append(captioner(sd_image_a))
+        caption.append(captioner(sd_image_a))
         #caption.append(captioner_2(sd_image_a))
-        caption.append(captioner_3(sd_image_a))
+        #caption.append(captioner_3(sd_image_a))
         caption_2.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)
-            #caption.append(captioner(sd_image_b))
+            caption.append(captioner(sd_image_b))
             #caption.append(captioner_2(sd_image_b))
-            caption.append(captioner_3(sd_image_b))
+            #caption.append(captioner_3(sd_image_b))
             caption_2.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)
-            #caption.append(captioner(sd_image_c))
+            caption.append(captioner(sd_image_c))
             #caption.append(captioner_2(sd_image_c))
-            caption.append(captioner_3(sd_image_c))
+            #caption.append(captioner_3(sd_image_c))
             caption_2.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)
-            #caption.append(captioner(sd_image_d))
+            caption.append(captioner(sd_image_d))
             #caption.append(captioner_2(sd_image_d))
-            caption.append(captioner_3(sd_image_d))
+            #caption.append(captioner_3(sd_image_d))
             caption_2.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)
-            #caption.append(captioner(sd_image_e))
+            caption.append(captioner(sd_image_e))
             #caption.append(captioner_2(sd_image_e))
-            caption.append(captioner_3(sd_image_e))
+            #caption.append(captioner_3(sd_image_e))
             caption_2.append(captioning(sd_image_e))
         else:
             sd_image_e = None