Model card: add the independent CUDA replication (random run order)

README.md

@@ -86,19 +86,22 @@ Use that loss to match the reference training.
 
 Fine-tuning on the 61-word task with the authors' `run_cls.py` protocol (200
 epochs, AdamW, sigmoid cross-entropy, 6 seeds × 12 subjects, paper-time data
-revision `056a87b7`, mean ± SE over subjects). The authors' fine-tuning loop
-was run with both the original model and the braindecode port, starting from
-the same weights:
+revision `056a87b7`, mean ± SE over subjects). There are two independent runs:
 
-| | paper (Table 2) | authors' code | braindecode port |
-|---|---|---|---|
-| Du-IN (mae) | 62.70 ± 4.69 | 60.52 ± 4.58 | 60.16 ± 4.62 |
-| Du-IN (vqvae) | 58.24 ± 4.83 | 55.10 ± 4.51 | 55.00 ± 4.51 |
-| scratch | 56.29 ± 5.20 | 54.06 ± 5.12 | 53.73 ± 5.13 |
+- **A**: the authors' `run_cls.train()` loop run with both the original model
+  and the port, starting from the same weights (Gaudi2 / CPU);
+- **B**: a standalone braindecode re-implementation of the same loop with the
+  port, on CUDA, with the run order drawn at random per seed (see below).
+
+| | paper (Table 2) | A: authors' code | A: port | B: port |
+|---|---|---|---|---|
+| Du-IN (mae) | 62.70 ± 4.69 | 60.52 ± 4.58 | 60.16 ± 4.62 | 59.37 ± 4.69 |
+| Du-IN (vqvae) | 58.24 ± 4.83 | 55.10 ± 4.51 | 55.00 ± 4.51 | 54.02 ± 4.57 |
+| scratch | 56.29 ± 5.20 | 54.06 ± 5.12 | 53.73 ± 5.13 | 53.77 ± 5.10 |
 
 The port matches the authors' code to within 0.4 points. Training the port on
 plain logits (`cross_entropy(logits)`) instead of the sigmoid loss gives 55.6
-for mae.
+for mae in A and 54.5 in B.
 
 **Split caveat.** The authors' split concatenates a subject's runs in
 `os.listdir` order and puts the first 80 % of each word's trials in the