Spaces:
Running
on
Zero
Running
on
Zero
Update app.py
Browse files
app.py
CHANGED
@@ -292,8 +292,8 @@ with gr.Blocks(css=css, theme="bethecloud/storj_theme") as demo:
|
|
292 |
)
|
293 |
gr.Markdown("**Model Info 💻** | [Report Bug](https://huggingface.co/spaces/prithivMLmods/Multimodal-OCR-Comparator/discussions)")
|
294 |
gr.Markdown("> Camel-Doc-OCR-062825 and Megalodon-OCR-Sync-0713 are both fine-tuned versions of the Qwen2.5-VL series focused on document retrieval, content extraction, analysis recognition, and excelling in OCR and visual document analysis tasks for structured and unstructured content—Camel-Doc-OCR-062825 leveraging the Qwen2.5-VL-7B-Instruct as its base, while Megalodon-OCR-Sync-0713 uses Qwen2.5-VL-3B-Instruct and is especially trained on diverse captioning datasets.")
|
295 |
-
gr.Markdown("GLM-4.1V-9B-Thinking is a vision-language model (VLM) based on the GLM-4-9B-0414 foundation, with a strong emphasis on advanced reasoning capabilities, chain-of-thought inference, and robust bilingual (Chinese/English) performance on complex multimodal benchmarks.")
|
296 |
-
gr.Markdown("DeepEyes-7B stands out for its agentic reinforcement learning approach, focusing on thinking with images for better visual reasoning, math problem-solving, and mitigating hallucination using Qwen2.5-VL-7B-Instruct as its foundation. Finally, Qwen2.5-VL-3B-Instruct-abliterated is part of the Qwen2.5-VL family, known for its versatile vision-language understanding and generation, serving as the foundational architecture for several of these fine-tuned vision-language and OCR models.")
|
297 |
|
298 |
# Define the submit button actions
|
299 |
image_submit.click(fn=generate_image,
|
|
|
292 |
)
|
293 |
gr.Markdown("**Model Info 💻** | [Report Bug](https://huggingface.co/spaces/prithivMLmods/Multimodal-OCR-Comparator/discussions)")
|
294 |
gr.Markdown("> Camel-Doc-OCR-062825 and Megalodon-OCR-Sync-0713 are both fine-tuned versions of the Qwen2.5-VL series focused on document retrieval, content extraction, analysis recognition, and excelling in OCR and visual document analysis tasks for structured and unstructured content—Camel-Doc-OCR-062825 leveraging the Qwen2.5-VL-7B-Instruct as its base, while Megalodon-OCR-Sync-0713 uses Qwen2.5-VL-3B-Instruct and is especially trained on diverse captioning datasets.")
|
295 |
+
gr.Markdown("> GLM-4.1V-9B-Thinking is a vision-language model (VLM) based on the GLM-4-9B-0414 foundation, with a strong emphasis on advanced reasoning capabilities, chain-of-thought inference, and robust bilingual (Chinese/English) performance on complex multimodal benchmarks.")
|
296 |
+
gr.Markdown("> DeepEyes-7B stands out for its agentic reinforcement learning approach, focusing on thinking with images for better visual reasoning, math problem-solving, and mitigating hallucination using Qwen2.5-VL-7B-Instruct as its foundation. Finally, Qwen2.5-VL-3B-Instruct-abliterated is part of the Qwen2.5-VL family, known for its versatile vision-language understanding and generation, serving as the foundational architecture for several of these fine-tuned vision-language and OCR models.")
|
297 |
|
298 |
# Define the submit button actions
|
299 |
image_submit.click(fn=generate_image,
|