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/ptrm_same_subset/paired_ptrm_k100_n1000_seed0_summary.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/ptrm_same_subset/paired_ptrm_k100_n1000_seed0_summary.csv')
| -rw-r--r-- | research/flossing/ptrm_same_subset/paired_ptrm_k100_n1000_seed0_summary.csv | 2 |
1 files changed, 2 insertions, 0 deletions
diff --git a/research/flossing/ptrm_same_subset/paired_ptrm_k100_n1000_seed0_summary.csv b/research/flossing/ptrm_same_subset/paired_ptrm_k100_n1000_seed0_summary.csv new file mode 100644 index 0000000..8122604 --- /dev/null +++ b/research/flossing/ptrm_same_subset/paired_ptrm_k100_n1000_seed0_summary.csv @@ -0,0 +1,2 @@ +n,rollouts,base_det,multi4_det,delta_det,base_mean_rollout,multi4_mean_rollout,delta_mean_rollout,base_qmax,multi4_qmax,delta_qmax,base_oracle,multi4_oracle,delta_oracle,base_correct_count_mean,multi4_correct_count_mean,delta_correct_count_mean,det_base_only_frac,det_multi4_only_frac,oracle_base_only_frac,oracle_multi4_only_frac +1000,100,0.887,0.911,0.02400000000000002,0.94188,0.95417,0.012289999999999912,0.984,0.988,0.0040000000000000036,0.985,0.988,0.0030000000000000027,94.188,95.417,1.2289999999999992,0.034,0.058,0.001,0.004 |
