Update app.py

app.py

@@ -45,11 +45,11 @@ MAX_INPUT_TOKEN_LENGTH = int(os.getenv("MAX_INPUT_TOKEN_LENGTH", "4096"))
 
 device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
 
-# Load RolmOCR
-MODEL_ID_M = "reducto/RolmOCR"
-processor_m = AutoProcessor.from_pretrained(MODEL_ID_M, trust_remote_code=True)
-model_m = Qwen2_5_VLForConditionalGeneration.from_pretrained(
-    MODEL_ID_M,
+# Load Nanonets-OCR-s
+MODEL_ID_V = "nanonets/Nanonets-OCR-s"
+processor_v = AutoProcessor.from_pretrained(MODEL_ID_V, trust_remote_code=True)
+model_v = Qwen2_5_VLForConditionalGeneration.from_pretrained(
+    MODEL_ID_V,
     trust_remote_code=True,
     torch_dtype=torch.float16
 ).to(device).eval()
@@ -63,15 +63,6 @@ model_x = Qwen2VLForConditionalGeneration.from_pretrained(
     torch_dtype=torch.float16
 ).to(device).eval()
 
-# Load Nanonets-OCR-s
-MODEL_ID_V = "nanonets/Nanonets-OCR-s"
-processor_v = AutoProcessor.from_pretrained(MODEL_ID_V, trust_remote_code=True)
-model_v = Qwen2_5_VLForConditionalGeneration.from_pretrained(
-    MODEL_ID_V,
-    trust_remote_code=True,
-    torch_dtype=torch.float16
-).to(device).eval()
-
 # Load Aya-Vision-8b
 MODEL_ID_A = "CohereForAI/aya-vision-8b"
 processor_a = AutoProcessor.from_pretrained(MODEL_ID_A, trust_remote_code=True)
@@ -81,17 +72,23 @@ model_a = AutoModelForImageTextToText.from_pretrained(
     torch_dtype=torch.float16
 ).to(device).eval()
 
-#-----------------------------subfolder-----------------------------#
-# Load MonkeyOCR-3B-0709
-MODEL_ID_W = "echo840/MonkeyOCR-3B-0709"
-SUBFOLDER = "Recognition"
-processor_w = AutoProcessor.from_pretrained(MODEL_ID_W, trust_remote_code=True, subfolder=SUBFOLDER)
+# Load AIGVE-MACS
+MODEL_ID_W = "xiaoliux/AIGVE-MACS"
+processor_w = AutoProcessor.from_pretrained(MODEL_ID_W, trust_remote_code=True)
 model_w = Qwen2_5_VLForConditionalGeneration.from_pretrained(
-    MODEL_ID_W, trust_remote_code=True,
-    subfolder=SUBFOLDER,
-    torch_dtype=torch.float16).to(device).eval()
-#-----------------------------subfolder-----------------------------#
+    MODEL_ID_W, 
+    trust_remote_code=True,
+    torch_dtype=torch.float16
+).to(device).eval()
 
+# Load RolmOCR
+MODEL_ID_M = "reducto/RolmOCR"
+processor_m = AutoProcessor.from_pretrained(MODEL_ID_M, trust_remote_code=True)
+model_m = Qwen2_5_VLForConditionalGeneration.from_pretrained(
+    MODEL_ID_M,
+    trust_remote_code=True,
+    torch_dtype=torch.float16
+).to(device).eval()
 
 def downsample_video(video_path):
     """
@@ -128,7 +125,7 @@ def generate_image(model_name: str, text: str, image: Image.Image,
     if model_name == "RolmOCR-7B":
         processor = processor_m
         model = model_m
-    elif model_name == "Qwen2-VL-OCR":
+    elif model_name == "Qwen2-VL-OCR-2B":
         processor = processor_x
         model = model_x
     elif model_name == "Nanonets-OCR-s":
@@ -137,7 +134,7 @@ def generate_image(model_name: str, text: str, image: Image.Image,
     elif model_name == "Aya-Vision-8B":
         processor = processor_a
         model = model_a
-    elif model_name == "MonkeyOCR-3B-0709":
+    elif model_name == "AIGVE-MACS-7B":
         processor = processor_w
         model = model_w
     else:
@@ -189,7 +186,7 @@ def generate_video(model_name: str, text: str, video_path: str,
     if model_name == "RolmOCR-7B":
         processor = processor_m
         model = model_m
-    elif model_name == "Qwen2-VL-OCR":
+    elif model_name == "Qwen2-VL-OCR-2B":
         processor = processor_x
         model = model_x
     elif model_name == "Nanonets-OCR-s":
@@ -198,7 +195,7 @@ def generate_video(model_name: str, text: str, video_path: str,
     elif model_name == "Aya-Vision-8B":
         processor = processor_a
         model = model_a
-    elif model_name == "MonkeyOCR-3B-0709":
+    elif model_name == "AIGVE-MACS-7B":
         processor = processor_w
         model = model_w
     else:
@@ -312,17 +309,16 @@ with gr.Blocks(css=css, theme="bethecloud/storj_theme") as demo:
                 with gr.Accordion("(Result.md)", open=False):
                     markdown_output = gr.Markdown(label="Formatted Result (Result.Md)")
             model_choice = gr.Radio(
-                choices=["Nanonets-OCR-s", "MonkeyOCR-3B-0709", "Qwen2-VL-OCR", "RolmOCR-7B", "Aya-Vision-8B"],
+                choices=["Nanonets-OCR-s", "Qwen2-VL-OCR-2B", "RolmOCR-7B", "AIGVE-MACS-7B", "Aya-Vision-8B"],
                 label="Select Model",
                 value="Nanonets-OCR-s"
             )
             gr.Markdown("**Model Info 💻** | [Report Bug](https://huggingface.co/spaces/prithivMLmods/Multimodal-OCR/discussions)")
             gr.Markdown("> [Nanonets-OCR-s](https://huggingface.co/nanonets/Nanonets-OCR-s): nanonets-ocr-s is a powerful, state-of-the-art image-to-markdown ocr model that goes far beyond traditional text extraction. it transforms documents into structured markdown with intelligent content recognition and semantic tagging.")
-            gr.Markdown("> [MonkeyOCR-3B-0709](https://huggingface.co/echo840/MonkeyOCR-3B-0709): MonkeyOCR adopts a structure-recognition-relation (SRR) triplet paradigm, which simplifies the multi-tool pipeline of modular approaches while avoiding the inefficiency of using large multimodal models for full-page document processing.")
-            gr.Markdown("> [Qwen2-VL-OCR-2B-Instruct](https://huggingface.co/prithivMLmods/Qwen2-VL-OCR-2B-Instruct): qwen2-vl-ocr-2b-instruct model is a fine-tuned version of qwen2-vl-2b-instruct, tailored for tasks that involve [messy] optical character recognition (ocr), image-to-text conversion, and math problem solving with latex formatting.")
+            gr.Markdown("> [AIGVE-MACS-7B](https://huggingface.co/xiaoliux/AIGVE-MACS): AIGVE-MACS is a unified Vision-Language Model (VLM) designed to evaluate AI-generated videos. It provides numerical scores (0–5) and natural language justifications across 9 human-aligned video quality aspects.")
+            gr.Markdown("> [Qwen2-VL-OCR-2B](https://huggingface.co/prithivMLmods/Qwen2-VL-OCR-2B-Instruct): qwen2-vl-ocr-2b-instruct model is a fine-tuned version of qwen2-vl-2b-instruct, tailored for tasks that involve [messy] optical character recognition (ocr), image-to-text conversion, and math problem solving with latex formatting.")
             gr.Markdown("> [RolmOCR](https://huggingface.co/reducto/RolmOCR): rolmocr, high-quality, openly available approach to parsing pdfs and other complex documents optical character recognition. it is designed to handle a wide range of document types, including scanned documents, handwritten text, and complex layouts.")
             gr.Markdown("> [Aya-Vision](https://huggingface.co/CohereLabs/aya-vision-8b): cohere labs aya vision 8b is an open weights research release of an 8-billion parameter model with advanced capabilities optimized for a variety of vision-language use cases, including ocr, captioning, visual reasoning, summarization, question answering, code, and more.")
-            
             gr.Markdown(">⚠️note: all the models in space are not guaranteed to perform well in video inference use cases.")
         
     image_submit.click(