Downloads · 30 days
12
23% of all-time downloads
Dreamer-VLA/dreamer-vla-lapa-aux-libero
dreamer-vla-lapa-aux-libero is a robotics model from Dreamer-VLA. Use it for the robotics task on the model card, and read the license before you ship it in a product. It is set up for pytorch. The card lists the license as cc-by-nc-4.0.
LAPA's LAQ encoder (laqopenx, 344M params) unfrozen and finetuned on LIBERO with the inverse-dynamics auxiliary loss as the only supervision. 30k steps, batch 2, single GPU. Parallels the V-JEPA+aux pipeline for the L…
Downloads · 30 days
12
23% of all-time downloads
All-time downloads
53
Public
Repo size
1.4 GB
Likes
0
Public
Click a slice to open those files.
.pt1.4 GB · 100%
From the Hugging Face model README
LAPA's LAQ encoder (laq_openx, 344M params) unfrozen and finetuned on LIBERO with the inverse-dynamics auxiliary loss as the only supervision. 30k steps, batch 2, single GPU. Parallels the V-JEPA+aux pipeline for the LAPA family in the dreamer-vla paper.
| Variant | params | train R² | test R² |
|---|---|---|---|
| LAPA frozen (reference) | 344M | +0.240 | +0.410 |
| LAPA + aux finetune ⭐ | 344M | +0.489 | +0.510 |
aux loss adds +0.10 over frozen baseline — comparable to V-JEPA 2 ViT-L frozen → +aux jump (+0.40 → +0.85, +0.44) but smaller magnitude.
laq_openx.pt (LAPA's pretrained LAQ on Open X-Embodiment)requires_grad=True)InvDynAuxHead(in_dim=1024, action_dim=7, hidden=[512,256])import sys, torch
sys.path.insert(0, "external_models/LAPA/laq")
from laq_model.latent_action_quantization import LatentActionQuantization
state = torch.load("ckpt_last.pt", weights_only=False)
model = LatentActionQuantization(
dim=1024, quant_dim=32, codebook_size=8,
image_size=256, patch_size=32,
spatial_depth=8, temporal_depth=8,
dim_head=64, heads=16, code_seq_len=4,
)
model.load_state_dict(state["encoder"])
The aux head is also in the ckpt (state["aux_head"]) but is training-only
and ignored at probe time.
@inproceedings{Ye2025LAPA,
title={Latent Action Pretraining from Videos},
author={Ye, Seonghyeon and others},
year={2024}, booktitle={NeurIPS}
}
(dreamer-vla paper citation when on arXiv.)