analysis/loss_curves_summary.md

Training loss curves — conclusions

Regenerated from wandb (scripts/plot_loss_curves.py). Runs covered: 37.

• shift16 + delta-to-ref are the two tactile fixes; without shift16 the tactile stream diverges.
• cam-pose curves look fine on val_loss but decoded LPIPS is far worse — see perceptual table.
• stride2 (4x data) is the biggest perceptual win; p01 trades a little val_loss for rollout consistency.
• hybrid_flow uses rectified flow — its val_loss is on a different scale, judge by decoded metrics.

Included runs: mm_v0, mot_vt_p01, mot_vta_p01, mv1_256_delta_shift, mv1_3v2t_delta_ref, mv1_p01, mv1_p3_delta_shift, mv1_v2_perframe, mv1_v2_stride2, mv1_v2_stride2_p01_adaptw, mv1_v2_stride2_p01_hybrid, mv1_v2_stride2_p01_hybrid_deep, mv1_v2_stride2_p01_hybrid_flow, mv1_v2_stride2_p01_hybrid_rope, mv1_v2_stride2_p01_plucker, mv1_v2_stride2_p01_tactile_only, mv1_v2_stride2_p01_v2t, mv1_v2_stride2_p01_vision_only, mv1_v2_stride2_p01rand, mv2_3v2t_campose_delta_ref, mv2_p4_delta_shift, p2_mv, p3_mv2t, p4_gate, p5_gate, vo_left, vo_middle, vo_right, vo_v0, vtwm_e2_delta_ref, vtwm_e3_delta_ref_null, vtwm_n1_shift16_delta_ref_null_prenoise, vtwm_n2_shift16_delta_ref_tactile_dropout, vtwm_p0_shift16_delta_episode_p01_ref, vtwm_s1_shift16_mmdit_mv, vtwm_s2_shift16_delta_ref, vtwm_s3_shift16_delta_ref_weight

Not on wandb yet (curve omitted): mv1_v2_h4f12, mv1_v2_stride2_p01, p4_mv2t_extr, vtwm_e1_1v2t_mmdit_mv