Search papers, labs, and topics across Lattice.
This study conducts the largest controlled scaling analysis of pixel-wise Earth-observation foundation models, involving 395 training runs on 1,024 GH200 superchips. The findings reveal that pretraining loss is a poor predictor of downstream performance, suggesting that traditional selection methods waste computational resources. By establishing a guideline for scaling鈥攚here the encoder and data should grow together while keeping the projector fixed鈥攖he authors successfully train a family of models that outperform larger counterparts, culminating in the release of the TESSERA v2-1B-M, which achieves high performance with significantly reduced storage requirements.
Pretraining loss is a misleading metric, leading to inefficient compute use, while a simple scaling rule can unlock superior performance in pixel-wise Earth-observation models.
Pixel-wise Earth-observation (EO) foundation models are now achieving state-of-the-art performance via generated spatial embeddings. However, how these models scale and how best to spend a pretraining budget remain poorly understood. We present the largest controlled scaling study for EO to date: 395 training runs on 1,024 GH200 superchips within a fixed pixel-wise Barlow Twins family, each evaluated on 15 downstream tasks. We find that pretraining loss barely predicts downstream performance (|Pearson r|<0.2), so selecting models by loss wastes a large share of the compute. We also find that, as the training budget grows, the encoder and the data should grow together while the projector stays fixed, which gives a simple rule for allocating compute. Using this rule, we train a family of pixel-wise models (0.5B and 1B, with a 2B model in training) and distill them into compact students for embeddings-as-data deployment. The 21-million-parameter distilled TESSERA v2-1B-M in aggregate outperforms all open and proprietary models tested, some of which are orders of magnitude larger. These students produce Matryoshka representations that are inexpensive to serve: a 16-dimensional prefix keeps 92% of the full 128-dimensional performance at 1/8 of the storage. Upon completion of training we plan to release v2 global embeddings covering 2017-2025. Together, these results give a concrete, empirically grounded recipe for scaling pixel-wise EO foundation models: train large encoders, select by downstream performance, and distil into flexible student models. All code will be released at https://github.com/ucam-eo/tessera.
