fix text

Author: Exiam6

cambrian-p/index.html

@@ -295,7 +295,7 @@ <h2>Pose-Grounded Video Understanding</h2>
                         </div>
                         <div class="icon-item">
                             <img src="./static/img/icons/training.svg" alt="Training Icon">
-                            <div><strong>Interleaved Training</strong>: Pose estimation and Video Query Answering disagree on how to sample and augment frames. We interleave pose-only batches with jittered VQA batches, balancing gradients from both objectives in a single training run.</div>
+                            <div><strong>Interleaved Training</strong>: Pose estimation and Video Query Answering normally have different strategies for how to sample and augment frames. We interleave pose-only batches with jittered VQA batches, balancing gradients from both objectives in a single training run.</div>
                         </div>
                         <div class="icon-item">
                             <img src="./static/img/icons/vqa.svg" alt="VQA Icon">