Update app.py

app.py

@@ -8,7 +8,7 @@ import spaces
 from PIL import Image
 import requests
 import transformers
-from transformers import T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 from translatepy import Translator
 
 os.environ["HF_HUB_ENABLE_HF_TRANSFER"] = "1"
@@ -54,9 +54,15 @@ text_encoder_3 = T5EncoderModel.from_pretrained(
     torch_dtype=torch.float16,
 )
 
+tokenizer_3 = AutoTokenizer.from_pretrained(
+    repo, 
+    subfolder="tokenizer_3", 
+    torch_dtype=torch.float16,
+)
+
 # Ensure model and scheduler are initialized in GPU-enabled function
 if torch.cuda.is_available():
-    pipe = StableDiffusion3Pipeline.from_pretrained(repo, vae=vae, transformer=transformer, text_encoder_3=text_encoder_3, torch_dtype=torch.float16).to("cuda")
+    pipe = StableDiffusion3Pipeline.from_pretrained(repo, vae=vae, transformer=transformer, tokenizer_3=tokenizer_3, text_encoder_3=text_encoder_3, torch_dtype=torch.float16).to("cuda")
 
 pipe.scheduler = FlowMatchEulerDiscreteScheduler.from_config(pipe.scheduler.config)
 
@@ -73,14 +79,23 @@ def generate_image(
 
     if seed == -1:
         seed = random.randint(0, MAX_SEED)
-    generator = torch.Generator().manual_seed(seed)
     
-    prompt = str(translator.translate(prompt, 'English'))
-
     print(f'prompt:{prompt}')
     
+    text = str(translator.translate(prompt['text'], 'English'))
+
+    
+    if prompt['files']:
+        images = Image.open(prompt['files'][-1]).convert('RGB')
+    else:
+        images = None
+    generator = torch.Generator().manual_seed(seed)
+    
+
+
     image = pipe(
-        prompt, 
+        text,
+        image=images,
         negative_prompt=negative, 
         width=width,
         height=height,
@@ -109,10 +124,10 @@ examples = [
 
 with gr.Blocks(css=CSS, js=JS, theme="soft") as demo:
     gr.HTML("<h1><center>SD3M🦄</center></h1>")
-    gr.HTML("<p><center><a href='https://huggingface.co/stabilityai/stable-diffusion-3-medium'>sd3m</a> text-to-image generation</center></p>")
+    gr.HTML("<p><center><a href='https://huggingface.co/stabilityai/stable-diffusion-3-medium'>sd3m</a> text/img-to-image generation</center></p>")
     with gr.Group():
         with gr.Row():
-            prompt = gr.Textbox(label='Enter Your Prompt (Multi-Languages)', value="best quality, HD, aesthetic", scale=6)
+            prompt = gr.MultimodalTextbox(label='Enter Your Prompt (Multi-Languages)', value="best quality, HD", file_types=['image'], scale=6)
             submit = gr.Button(scale=1, variant='primary')
     img = gr.Image(label='SD3M Generated Image')
     with gr.Accordion("Advanced Options", open=False):