Update README.md

README.md

@@ -9,7 +9,6 @@ base_model:
 ---
 
 # SWE-Star-32B
-
 ## Introduction
 
 [SWE-Star](https://huggingface.co/collections/LogicStar/swe-star) is a family of language models based on the [Qwen2.5-Coder](https://huggingface.co/collections/Qwen/qwen25-coder) family and trained on the [SWE-Star](https://huggingface.co/datasets/LogicStar/SWE-Star) dataset. The dataset contains approximately 250k agentic coding trajectories distilled from [Devstral-2-Small](https://huggingface.co/mistralai/Devstral-Small-2-24B-Instruct-2512) using [SWE-Smith](https://swesmith.com/) tasks.
@@ -37,7 +36,7 @@ It is worth noting that the agent was run on [MN5](https://www.bsc.es/marenostru
 Our models also achieve very high Pass@16 rates, making them strong candidates for further reinforcement learning. Our 32B model reaches a Pass@16 score of 75.5%:
 
 <picture>
-  <source srcset="https://cdn-uploads.huggingface.co/production/uploads/689f18d4da73199a8848954b/9mxYRYcmicls_3fSoR__B.png" media="(prefers-color-scheme: dark)">
-  <source srcset="https://cdn-uploads.huggingface.co/production/uploads/689f18d4da73199a8848954b/XR-t5Jbtv_AWTkZSDR8fn.png" media="(prefers-color-scheme: light)">
-  <img src="https://cdn-uploads.huggingface.co/production/uploads/689f18d4da73199a8848954b/XR-t5Jbtv_AWTkZSDR8fn.png" width="1000">
+  <source srcset="https://cdn-uploads.huggingface.co/production/uploads/689f18d4da73199a8848954b/s6wolafS7nAEWnXMj_Sw5.png" media="(prefers-color-scheme: dark)">
+  <source srcset="https://cdn-uploads.huggingface.co/production/uploads/689f18d4da73199a8848954b/0uaPi5mUsuNnyCksENTvw.png" media="(prefers-color-scheme: light)">
+  <img src="https://cdn-uploads.huggingface.co/production/uploads/689f18d4da73199a8848954b/0uaPi5mUsuNnyCksENTvw.png" width="1000">
 </picture>