Update app.py

app.py

@@ -65,6 +65,15 @@ model_o = Qwen2VLForConditionalGeneration.from_pretrained(
     torch_dtype=torch.float16
 ).to(device).eval()
 
+# Load typhoon-ocr-3b
+MODEL_ID_J = "prithivMLmods/Megalodon-OCR-Sync-0713"
+processor_j = AutoProcessor.from_pretrained(MODEL_ID_J, trust_remote_code=True)
+model_j = Qwen2_5_VLForConditionalGeneration.from_pretrained(
+    MODEL_ID_J,
+    trust_remote_code=True,
+    torch_dtype=torch.float16
+).to(device).eval()
+
 def downsample_video(video_path):
     """
     Downsamples the video to evenly spaced frames.
@@ -108,6 +117,9 @@ def generate_image(model_name: str, text: str, image: Image.Image,
     elif model_name == "Typhoon-OCR-3B":
         processor = processor_t
         model = model_t
+    elif model_name == "Megalodon-OCR":
+        processor = processor_j
+        model = model_j
     else:
         yield "Invalid model selected.", "Invalid model selected."
         return
@@ -164,6 +176,9 @@ def generate_video(model_name: str, text: str, video_path: str,
     elif model_name == "Typhoon-OCR-3B":
         processor = processor_t
         model = model_t
+    elif model_name == "Megalodon-OCR":
+        processor = processor_j
+        model = model_j
     else:
         yield "Invalid model selected.", "Invalid model selected."
         return
@@ -281,20 +296,21 @@ with gr.Blocks(css=css, theme="bethecloud/storj_theme") as demo:
         with gr.Column():
             with gr.Column(elem_classes="canvas-output"):
                 gr.Markdown("## Output")
-                output = gr.Textbox(label="Raw Output Stream", interactive=False, lines=2)
+                output = gr.Textbox(label="Raw Output Stream", interactive=False, lines=2, show_copy_button=True)
 
                 with gr.Accordion("(Result.md)", open=False):                
                     markdown_output = gr.Markdown(label="(Result.Md)")
                 #download_btn = gr.Button("Download Result.md"
                 
             model_choice = gr.Radio(
-                choices=["DREX-062225-7B-exp", "olmOCR-7B-0225-preview", "VIREX-062225-7B-exp", "Typhoon-OCR-3B"],
+                choices=["DREX-062225-7B-exp", "Megalodon-OCR", "olmOCR-7B-0225-preview", "VIREX-062225-7B-exp", "Typhoon-OCR-3B"],
                 label="Select Model",
                 value="DREX-062225-7B-exp"
             )
 
             gr.Markdown("**Model Info 💻** | [Report Bug](https://huggingface.co/spaces/prithivMLmods/Doc-VLMs/discussions)")          
             gr.Markdown("> [DREX-062225-7B-exp](https://huggingface.co/prithivMLmods/DREX-062225-exp): the drex-062225-exp (document retrieval and extraction expert) model is a specialized fine-tuned version of docscopeocr-7b-050425-exp, optimized for document retrieval, content extraction, and analysis recognition. built on top of the qwen2.5-vl architecture.")
+            gr.Markdown("> [Megalodon-OCR-Sync-0713](https://huggingface.co/prithivMLmods/Megalodon-OCR-Sync-0713): megalodon-ocr-sync-0713 model is a fine-tuned version of qwen2.5-vl-3b-instruct, optimized for document retrieval, content extraction, and analysis recognition. built on top of the qwen2.5-vl architecture, this model enhances document comprehension capabilities.")
             gr.Markdown("> [VIREX-062225-7B-exp](https://huggingface.co/prithivMLmods/VIREX-062225-exp): the virex-062225-exp (video information retrieval and extraction expert - experimental) model is a fine-tuned version of qwen2.5-vl-7b-instruct, specifically optimized for advanced video understanding, image comprehension, sense of reasoning, and natural language decision-making through cot reasoning.")
             gr.Markdown("> [Typhoon-OCR-3B](https://huggingface.co/scb10x/typhoon-ocr-3b): a bilingual document parsing model built specifically for real-world documents in thai and english, inspired by models like olmocr, based on qwen2.5-vl-instruction. this model is intended to be used with a specific prompt only.")
             gr.Markdown("> [olmOCR-7B-0225](https://huggingface.co/allenai/olmOCR-7B-0225-preview): the olmocr-7b-0225-preview model is based on qwen2-vl-7b, optimized for document-level optical character recognition (ocr), long-context vision-language understanding, and accurate image-to-text conversion with mathematical latex formatting. designed with a focus on high-fidelity visual-textual comprehension.")