Today
-
piper_real3 OpenVLA control (INIT=none, LATENT=none), step 6000
Yesterday
-
root card: stack_bowls section (floor 0.667 deg), the new control, and reflow the real3 row into the 4-column table
-
stack_bowls_openvla_delta: matched control for stack_bowls (stock openvla-7b, no latent) + card
-
piper_real3_anchored_fwd step 6000
-
piper_real3_anchored_fwd step 5500
-
piper_real3_anchored_fwd step 5000
-
piper_real3_anchored_fwd step 4500
-
piper_real3_anchored_fwd step 4000
-
piper_real3_anchored_fwd step 3500
-
piper_real3_anchored_fwd step 3000
-
piper_real3_anchored_fwd step 2500
-
piper_real3_anchored_fwd step 2000
-
piper_real3_anchored_fwd step 1500
-
piper_real3_anchored_fwd step 1000
-
root README: piper_real3_anchored_fwd row
2 days ago
-
README: merge back the three-action-space section and the anchored exec-N numbers (do not overwrite this file)
-
LIBERO cells: record the phase-2 init explicitly in ctrl_meta, final cards
-
README: TERRA-vs-control inventory, and the 4 LIBERO cells
-
libero_10_frac10_terra_p2fft20k: TERRA on p2fft_20k, Long @10%, 60.6%
-
libero_goal_frac10_none: CONTROL no latent, Goal @10%, 88.6%
-
libero_goal_frac10_univla_cont: BASELINE UniVLA latent, Goal @10%, 89.4%
-
libero_goal_frac10_terra: TERRA latent, Goal @10%, 86.8%
-
READMEs: step 5500 numbers, fwd step 1500 row
-
press_button anchored, fixed window, step 1500
-
README: one table per model, which to test, and the two deployment traps
-
piper_pick_can_10k: record action_space=absolute explicitly
-
piper_pick_can_openvla_delta: merged delta policy + README
-
piper_pick_can_terra_delta: merged delta policy + README
-
READMEs: model catalogue at the root, one README per directory
-
chunk-anchored multi-task step 5500 (latest)
-
experimental chunk-anchored multi-task step 2500
3 days ago
-
root README: press_button_openvla_delta control row
-
press_button_terra_delta README: matched control result
-
press_button OpenVLA control (delta): 1.92 deg vs TERRA 0.92, ignores the instruction
-
press_button_terra_delta README: ensembling undecided for delta
-
README: add press_button_terra_delta row
-
press_button_terra_delta: TERRA arm, delta actions, 3000 steps (README: delta interface, held-out + lang-swap)
4 days ago
-
Handoff: the base is public now, no token needed
-
Card: document p2fft_20k, the phase-2 base, and that action_decoder.pt is required
-
Phase-2 pretrained base (OpenVLA-7B, 20k full-parameter steps, latp2_inv latent) as a post-training INIT
-
Handoff: clone URL, the settled padding answer, confirmed task names and episode lengths
-
Card: the real code URL (owner was missing, sent a reader to a 404) and drop the wrong `private: true`
-
Handoff brief: training the press_button arm on a Jetstream2 g3.xl
6 days ago
-
piper README: null baselines -- the policy beats the trajectory baselines 3-4x and loses to persistence, so do not rank it by action L1
-
piper README: the held-out number that matters is the open-loop one (3.24 deg / 0.149), not the oracle-conditioned 0.093
-
README: the repo now also holds a real-robot policy
-
piper_pick_can_10k: interface, training budget, held-out numbers, and what is not established
-
piper_pick_can_10k: 20k phase-2 base + LoRA/flow/q4, 10k steps on the Piper pick-can data
16 days ago
-
Add libero_goal_full
-
Add libero_10_full
