Improve model card: Add library name, pipeline tag, links, and base model

README.md

@@ -1,9 +1,10 @@
 ---
-license: apache-2.0
+base_model: lmms-lab/llava-onevision-qwen2-7b-ov
 datasets:
 - lmms-lab/EgoLife
-base_model:
-- lmms-lab/llava-onevision-qwen2-7b-ov
+license: apache-2.0
+library_name: transformers
+pipeline_tag: video-text-to-text
 tags:
 - multimodal
 ---
@@ -20,8 +21,8 @@ EgoGPT excels in two primary scenarios:
 
 
 For further details, please refer to the following resources:
-- 📰 Paper: https://arxiv.org/abs/2503.03803
-- 🪐 Project Page: https://github.com/EvolvingLMMs-Lab/EgoLife
+- 📰 Paper: [EgoLife: Towards Egocentric Life Assistant](https://arxiv.org/abs/2503.03803)
+- 🪐 Project Page: [https://egolife-ai.github.io/](https://egolife-ai.github.io/)
 - 📦 Datasets: https://huggingface.co/datasets/lmms-lab/EgoIT-99K & https://huggingface.co/datasets/lmms-lab/EgoLife
 - 🤗 Model Collections: https://huggingface.co/collections/lmms-lab/egolife-67c04574c2a9b64ab312c342
 
@@ -141,7 +142,10 @@ def main(
     model.eval()
 
     conv_template = "qwen_1_5"
-    question = f"<image>\n<speech>\n\n{query}"
+    question = f"<image>
+<speech>
+
+{query}"
     conv = copy.deepcopy(conv_templates[conv_template])
     conv.append_message(conv.roles[0], question)
     conv.append_message(conv.roles[1], None)