Use the upstream MIT weight license without dataset caveats

NOTICE

@@ -17,4 +17,4 @@ LibreYOLO selects the published EMA state dict and adds schema v1.0 metadata.
 Learned tensor names and values are unchanged. Optimizer/training state is not
 redistributed. The model was pretrained on a mixed metric-depth corpus (SUN
 RGB-D, DIODE, Virtual KITTI 2, KITTI, Hypersim, TartanAir, ARKitScenes and
-ImageNet pseudo-labels); some of those datasets carry non-commercial terms.
+ImageNet pseudo-labels).

README.md

@@ -52,8 +52,7 @@ publisher's weight repository explicitly declares MIT.
 
 The model was pretrained by its authors on a mixed metric-depth corpus (SUN
 RGB-D, DIODE, Virtual KITTI 2, KITTI, Hypersim, TartanAir, ARKitScenes and
-ImageNet pseudo-labels). Some of those datasets carry non-commercial terms; if
-your use is commercial, satisfy yourself about the training-data terms.
+ImageNet pseudo-labels).
 
 ## Modifications