Fix scorecard URLs.

README.md

@@ -74,7 +74,7 @@ from 1979 to 2018, and can be causally evaluated on 2019 and later years.
 This model has the smallest memory footprint of those provided and has been
 provided to enable low cost demonstrations (for example, it is runnable in a free Colab notebook).
 While its performance is reasonable, it is not representative of the performance
-of the GenCast models (1-3) above. For reference, a scorecard comparing its performance to ENS can be found in [docs/](https://github.com/google-deepmind/graphcast/docs/GenCast_1p0deg_Mini_ENS_scorecard.png). Note that in this scorecard,
+of the GenCast models (1-3) above. For reference, a scorecard comparing its performance to ENS can be found in [docs/](https://github.com/google-deepmind/graphcast/blob/main/docs/GenCast_1p0deg_Mini_ENS_scorecard.png). Note that in this scorecard,
 GenCast Mini only uses 8 member ensembles (vs. ENS' 50) so we use the fair (unbiased)
 CRPS to allow for fair comparison.
 

docs/cloud_vm_setup.md

@@ -68,7 +68,7 @@ This document describes how to run `gencast_demo_cloud_vm.ipynb` through [Colabo
   this originates from the attention mechanisms being algebraically equivalent, but
   not numerically equivalent. For reference, a scorecard comparing GenCast
   forecasts produced on a TPUv4 with `splash attention` vs. on a H100 with `triblockdiag_mha`
-  can be found in [docs/](https://github.com/google-deepmind/graphcast/docs/GenCast_0p25deg_accelerator_scorecard.png).
+  can be found in [docs/](https://github.com/google-deepmind/graphcast/blob/main/docs/GenCast_0p25deg_accelerator_scorecard.png).
   Note that this scorecard differs from those found in the GenCast paper
   in a number of ways:
     - 8 member ensembles (vs. 50 in the paper)