diff options
| author | YurenHao0426 <blackhao0426@gmail.com> | 2026-06-13 12:35:36 -0500 |
|---|---|---|
| committer | YurenHao0426 <blackhao0426@gmail.com> | 2026-06-13 12:35:36 -0500 |
| commit | 66e0d8b9fd4d0f7a2231d689c055e26fdf1cf04a (patch) | |
| tree | c29cba61124018755a19b02c9d33e3ad5f2e05cc /research/flossing/multi4_eval_compare/trm_baseline_eval.csv | |
Curated export for clone-and-run Maze training (2x A6000) + diagnostics.
trm/hrm pretrain.py carry trajectory-augmentation code (backward-compatible).
Heavy artifacts (checkpoints/wandb/npz) gitignored; see PROVENANCE.md.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Diffstat (limited to 'research/flossing/multi4_eval_compare/trm_baseline_eval.csv')
| -rw-r--r-- | research/flossing/multi4_eval_compare/trm_baseline_eval.csv | 11 |
1 files changed, 11 insertions, 0 deletions
diff --git a/research/flossing/multi4_eval_compare/trm_baseline_eval.csv b/research/flossing/multi4_eval_compare/trm_baseline_eval.csv new file mode 100644 index 0000000..0b6bc11 --- /dev/null +++ b/research/flossing/multi4_eval_compare/trm_baseline_eval.csv @@ -0,0 +1,11 @@ +run,step,all/exact_accuracy,all/accuracy,all/lm_loss,all/q_halt_accuracy,all/q_halt_loss,all/steps
+trm_baseline,26041,0.5575894117355347,0.8469749093055725,0.35285070538520813,0.9997445344924927,0.003797376062721014,16
+trm_baseline,52082,0.6295099854469299,0.8704391121864319,0.2973524034023285,0.9998486042022705,0.0011324305087327957,16
+trm_baseline,78123,0.6993892788887024,0.8920264840126038,0.2477506548166275,0.9998935461044312,0.0006260558147914708,16
+trm_baseline,104164,0.72928386926651,0.9020143151283264,0.22608688473701477,0.9998770356178284,0.001141014276072383,16
+trm_baseline,130205,0.7653990387916565,0.914251446723938,0.19878524541854858,0.999917209148407,0.0010435190051794052,16
+trm_baseline,156246,0.7596656680107117,0.9119072556495667,0.2041437327861786,0.999862790107727,0.0011596218682825565,16
+trm_baseline,182287,0.7541900873184204,0.9094774723052979,0.2100999504327774,0.9998249411582947,0.0013093978632241488,16
+trm_baseline,208328,0.7732800841331482,0.9166732430458069,0.19410833716392517,0.9998320937156677,0.0033004307188093662,16
+trm_baseline,234369,0.7750374674797058,0.9172152280807495,0.19279460608959198,0.9998533725738525,0.0032994903158396482,16
+trm_baseline,260410,0.7742025256156921,0.9169425964355469,0.19348150491714478,0.9998462796211243,0.002894919365644455,16
|