blumenstiel commited on
Commit
caa3514
·
1 Parent(s): 687a219

Update README

Browse files
Files changed (1) hide show
  1. README.md +2 -1
README.md CHANGED
@@ -23,7 +23,7 @@ Samples from the TerraMesh dataset with seven spatiotemporal aligned modalities.
23
 
24
  ## Dataset organisation
25
 
26
- The archive ships two top‑level splits `train/` and `val/`, each holding one folder per modality. More details with the dataset release end of June.
27
 
28
  ---
29
 
@@ -46,6 +46,7 @@ TerraMesh was used to pre-train [TerraMind-B](https://huggingface.co/ibm-esa-geo
46
  On the six evaluated segmentation tasks from PANGAEA bench, TerraMind‑B reaches an average mIoU of 66.6%, the best overall score with an average rank of 2.33. This amounts to roughly a 3pp improvement over the next‑best open model (CROMA), underscoring the benefits of pre‑training on TerraMesh.
47
  Compared to an ablation model pre-trained only on SSL4EO-S12 locations TerraMind-B performs overall 1pp better with better global generalization on more remote tasks like CTM-SS.
48
  More details in our [paper](https://arxiv.org/abs/2504.11172).
 
49
  ---
50
 
51
  ## Citation
 
23
 
24
  ## Dataset organisation
25
 
26
+ The archive ships two top‑level splits `train/` and `val/`, each holding one folder per modality. More details follow with the dataset release end of June.
27
 
28
  ---
29
 
 
46
  On the six evaluated segmentation tasks from PANGAEA bench, TerraMind‑B reaches an average mIoU of 66.6%, the best overall score with an average rank of 2.33. This amounts to roughly a 3pp improvement over the next‑best open model (CROMA), underscoring the benefits of pre‑training on TerraMesh.
47
  Compared to an ablation model pre-trained only on SSL4EO-S12 locations TerraMind-B performs overall 1pp better with better global generalization on more remote tasks like CTM-SS.
48
  More details in our [paper](https://arxiv.org/abs/2504.11172).
49
+
50
  ---
51
 
52
  ## Citation