spikefly/SemanticVLA-LIBERO

SemanticVLA · LIBERO

🚧 Placeholder. The URL is stable; checkpoints will be uploaded incrementally per the release roadmap.

SemanticVLA finetuned on the LIBERO benchmark.

Headline result

Suite	Success rate
libero_spatial	0.988
libero_object	0.996
libero_goal	0.986
libero_10	0.966
4-suite mean	0.9840

Best configuration: TL_saembs_lw010 (trace + latent semantic output, sa_embs injection, LM loss weight 0.10, step 30000).

Architecture

• Backbone: Qwen2.5-VL-3B (with trace + latent-action semantic heads)
• Action head: GR00T-style flow-matching expert (continuous action chunks)
• LAM tokenizer: SemanticVLA-LAM → libero/v5
• Action horizon: 16

Planned layout

SemanticVLA-LIBERO/
├── tl-saembs-lw010-best/
│   ├── pytorch_model.pt
│   ├── config.yaml
│   └── model_card.md
└── README.md

Additional ablation variants (L_none_lw010, TL_none_lw010, TL_saembs_lw005, etc.) may be uploaded as additional subdirectories upon release.

Sibling SemanticVLA checkpoint repos

Repo	Purpose
🤗 SemanticVLA-LAM	LAM tokenizers used by this VLA
🤗 SemanticVLA-Bridge	Bridge-finetuned VLA for SimplerEnv WidowX

Related resources

• Code: https://github.com/Fei-Ni/SemanticVLA_Offcial
• Datasets: https://hf.co/collections/spikefly/semanticvla-datasets
• Collection · Model Zoo: https://hf.co/collections/spikefly/semanticvla-model-zoo

How to load (placeholder API)

from huggingface_hub import hf_hub_download
import torch

ckpt = hf_hub_download(
    repo_id="spikefly/SemanticVLA-LIBERO",
    filename="tl-saembs-lw010-best/pytorch_model.pt",
)
state = torch.load(ckpt, map_location="cpu")
# loader will be released with the code repo

Citation

@misc{semanticvla2026,
  title  = {SemanticVLA: Semantic Vision-Language-Action Models with Trace-Conditioned Latent Action Tokens},
  author = {SemanticVLA Authors},
  year   = {2026},
  url    = {https://github.com/Fei-Ni/SemanticVLA_Offcial}
}

@inproceedings{liu2023libero,
  title     = {LIBERO: Benchmarking Knowledge Transfer for Lifelong Robot Learning},
  author    = {Liu, Bo and Zhu, Yifeng and others},
  booktitle = {NeurIPS Datasets and Benchmarks},
  year      = {2023}
}

License

Released under the MIT License.