Update app.py

app.py

@@ -1,5 +1,6 @@
 import os
 import gradio as gr
+from gradio_client import Client, handle_file
 import torch
 import spaces
 from diffusers import Lumina2Pipeline
@@ -12,14 +13,14 @@ else:
 
 # Load models
 def load_models():
-    model_name = "X-ART/LeX-Enhancer-full"
+    # model_name = "X-ART/LeX-Enhancer-full"
     
-    model = AutoModelForCausalLM.from_pretrained(
-        model_name,
-        torch_dtype=torch.bfloat16,
-        device_map="auto"
-    )
-    tokenizer = AutoTokenizer.from_pretrained(model_name)
+    # model = AutoModelForCausalLM.from_pretrained(
+    #     model_name,
+    #     torch_dtype=torch.bfloat16,
+    #     device_map="auto"
+    # )
+    # tokenizer = AutoTokenizer.from_pretrained(model_name)
     
     pipe = Lumina2Pipeline.from_pretrained(
         "X-ART/LeX-Lumina", 
@@ -28,9 +29,10 @@ def load_models():
     device = "cuda" if torch.cuda.is_available() else "cpu"
     pipe.to("cuda")
     
-    return model, tokenizer, pipe
+    return pipe
 
-model, tokenizer, pipe = load_models()
+pipe = load_models()
+client = Client("stzhao/LeX-Enhancer")
 
 def truncate_caption_by_tokens(caption, max_tokens=256):
     """Truncate the caption to fit within the max token limit"""
@@ -41,39 +43,39 @@ def truncate_caption_by_tokens(caption, max_tokens=256):
         print(f"Caption was truncated from {len(tokens)} tokens to {max_tokens} tokens")
     return caption
 
-@spaces.GPU(duration=70)
-def generate_enhanced_caption(image_caption, text_caption):
-    # model.to("cuda")
-    """Generate enhanced caption using the LeX-Enhancer model"""
-    combined_caption = f"{image_caption}, with the text on it: {text_caption}."
-    instruction = """
-Below is the simple caption of an image with text. Please deduce the detailed description of the image based on this simple caption. Note: 1. The description should only include visual elements and should not contain any extended meanings. 2. The visual elements should be as rich as possible, such as the main objects in the image, their respective attributes, the spatial relationships between the objects, lighting and shadows, color style, any text in the image and its style, etc. 3. The output description should be a single paragraph and should not be structured. 4. The description should avoid certain situations, such as pure white or black backgrounds, blurry text, excessive rendering of text, or harsh visual styles. 5. The detailed caption should be human readable and fluent. 6. Avoid using vague expressions such as "may be" or "might be"; the generated caption must be in a definitive, narrative tone. 7. Do not use negative sentence structures, such as "there is nothing in the image," etc. The entire caption should directly describe the content of the image. 8. The entire output should be limited to 200 words.
-"""
-    messages = [
-        {"role": "system", "content": "You are Qwen, created by Alibaba Cloud. You are a helpful assistant."},
-        {"role": "user", "content": instruction + "\nSimple Caption:\n" + combined_caption}
-    ]
-    text = tokenizer.apply_chat_template(
-        messages,
-        tokenize=False,
-        add_generation_prompt=True
-    )
-    model_inputs = tokenizer([text], return_tensors="pt").to(model.device)
+# @spaces.GPU(duration=70)
+# def generate_enhanced_caption(image_caption, text_caption):
+#     # model.to("cuda")
+#     """Generate enhanced caption using the LeX-Enhancer model"""
+#     combined_caption = f"{image_caption}, with the text on it: {text_caption}."
+#     instruction = """
+# Below is the simple caption of an image with text. Please deduce the detailed description of the image based on this simple caption. Note: 1. The description should only include visual elements and should not contain any extended meanings. 2. The visual elements should be as rich as possible, such as the main objects in the image, their respective attributes, the spatial relationships between the objects, lighting and shadows, color style, any text in the image and its style, etc. 3. The output description should be a single paragraph and should not be structured. 4. The description should avoid certain situations, such as pure white or black backgrounds, blurry text, excessive rendering of text, or harsh visual styles. 5. The detailed caption should be human readable and fluent. 6. Avoid using vague expressions such as "may be" or "might be"; the generated caption must be in a definitive, narrative tone. 7. Do not use negative sentence structures, such as "there is nothing in the image," etc. The entire caption should directly describe the content of the image. 8. The entire output should be limited to 200 words.
+# """
+#     messages = [
+#         {"role": "system", "content": "You are Qwen, created by Alibaba Cloud. You are a helpful assistant."},
+#         {"role": "user", "content": instruction + "\nSimple Caption:\n" + combined_caption}
+#     ]
+#     text = tokenizer.apply_chat_template(
+#         messages,
+#         tokenize=False,
+#         add_generation_prompt=True
+#     )
+#     model_inputs = tokenizer([text], return_tensors="pt").to(model.device)
 
-    generated_ids = model.generate(
-        **model_inputs,
-        max_new_tokens=1024
-    )
-    generated_ids = [
-        output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
-    ]
+#     generated_ids = model.generate(
+#         **model_inputs,
+#         max_new_tokens=1024
+#     )
+#     generated_ids = [
+#         output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
+#     ]
 
-    response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
-    enhanced_caption = response.split("</think>", -1)[-1].strip(" ").strip("\n")
-    model.to("cpu")
-    torch.cuda.empty_cache()
+#     response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
+#     enhanced_caption = response.split("</think>", -1)[-1].strip(" ").strip("\n")
+#     model.to("cpu")
+#     torch.cuda.empty_cache()
     
-    return combined_caption, enhanced_caption
+#     return combined_caption, enhanced_caption
 
 @spaces.GPU(duration=60)
 def generate_image(enhanced_caption, seed, num_inference_steps, guidance_scale):
@@ -113,7 +115,8 @@ def run_pipeline(image_caption, text_caption, seed, num_inference_steps, guidanc
     combined_caption = f"{image_caption}, with the text on it: {text_caption}."
     
     if enable_enhancer:
-        combined_caption, enhanced_caption = generate_enhanced_caption(image_caption, text_caption)
+        # combined_caption, enhanced_caption = generate_enhanced_caption(image_caption, text_caption)
+        combined_caption, enhanced_caption = client.predict(image_caption, text_caption)
     else:
         enhanced_caption = combined_caption