Fix pipeline tag, correct metadata typo, and add paper link

README.md

@@ -1,18 +1,17 @@
 ---
 base_model:
 - Qwen/Qwen2-VL-2B-Instruct
+datasets:
+- Tami3/HazardQA
+language:
+- en
 library_name: transformers
+license: apache-2.0
 model_name: HazardNet-unsloth-v0.4
+pipeline_tag: image-text-to-text
 tags:
 - trl
 - sft
-licence: license
-license: apache-2.0
-datasets:
-- Tami3/HazardQA
-language:
-- en
-pipeline_tag: visual-question-answering
 ---
 
 # Model Card for HazardNet-unsloth-v0.4
@@ -20,6 +19,8 @@ pipeline_tag: visual-question-answering
 This model is a fine-tuned version of [Qwen/Qwen2-VL-2B-Instruct](https://huggingface.co/Qwen/Qwen2-VL-2B-Instruct).
 It has been trained using [TRL](https://github.com/huggingface/trl).
 
+The model was presented in the paper [](https://hf.co/papers/2502.20572).
+
 ## Quick start
 
 ```python
@@ -30,7 +31,7 @@ from io import BytesIO
 
 # Initialize the Visual Question Answering pipeline with HazardNet
 hazard_vqa = pipeline(
-    "visual-question-answering",
+    "image-text-to-text",
     model="Tami3/HazardNet"
 )
 
@@ -82,7 +83,6 @@ except Exception as e:
 
  
 
-
 This model was trained with SFT.
 
 ### Framework versions
@@ -95,8 +95,6 @@ This model was trained with SFT.
 
 ## Citations
 
-
-
 Cite TRL as:
     
 ```bibtex