Update app.py

app.py

@@ -12,7 +12,7 @@ import torchvision.transforms.functional as TVF
 
 
 CLIP_PATH = "google/siglip-so400m-patch14-384"
-MODEL_PATH = "Qwen/Qwen2.5-7B-Instruct"
+MODEL_PATH = "meta-llama/Meta-Llama-3.1-8B"
 CHECKPOINT_PATH = Path("9em124t2-499968")
 TITLE = "<h1><center>JoyCaption Alpha One (2024-09-20a)</center></h1>"
 CAPTION_TYPE_MAP = {
@@ -138,7 +138,7 @@ text_model.eval()
 # Image Adapter
 print("Loading image adapter")
 image_adapter = ImageAdapter(clip_model.config.hidden_size, text_model.config.hidden_size, False, False, 38, False)
-image_adapter.load_state_dict(torch.load(CHECKPOINT_PATH / "image_adapter.pt", map_location="cpu"))
+image_adapter.load_state_dict(torch.load(CHECKPOINT_PATH / "image_adapter.pt", map_location="cpu", weights_only=True))
 image_adapter.eval()
 image_adapter.to("cuda")
 
@@ -375,7 +375,7 @@ with gr.Blocks(theme="Hev832/Applio", css=css) as demo:
 
             1. **Pick a Picture**: Find a cool picture you want to talk about and upload it.
 
-            2. **Choose What You Want**:
+            2. **Choose What You Want**: 
                - **Caption Type**: 
                  * "Descriptive" tells you what's in the picture
                  * "Training Prompt" helps computers make similar pictures
@@ -476,7 +476,7 @@ with gr.Blocks(theme="Hev832/Applio", css=css) as demo:
                 gr.Markdown("**Note:** Caption tone doesn't affect `rng-tags`, `training_prompt`, and `style_prompt`.")
 
                 run_button = gr.Button("Make My Caption!")
-            
+        
 
     def update_style_options(caption_type):
         return {
@@ -492,5 +492,6 @@ with gr.Blocks(theme="Hev832/Applio", css=css) as demo:
     
     run_button.click(fn=stream_chat, inputs=[input_image, caption_type, caption_tone, caption_length, lens_type, film_stock, composition_style, lighting_aspect, special_technique, color_effect], outputs=[output_caption])
 
+
 if __name__ == "__main__":
     demo.launch()