Don't Run with Scissors: Pruning Breaks VLA Models but They Can Be Recovered
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866914084326211584 |
|---|---|
| author | Jabbour, Jason Kim, Dong-Ki Smith, Max Patrikar, Jay Ghosal, Radhika Wang, Youhui Agha, Ali Reddi, Vijay Janapa Omidshafiei, Shayegan |
| author_facet | Jabbour, Jason Kim, Dong-Ki Smith, Max Patrikar, Jay Ghosal, Radhika Wang, Youhui Agha, Ali Reddi, Vijay Janapa Omidshafiei, Shayegan |
| contents | Vision-Language-Action (VLA) models have advanced robotic capabilities but remain challenging to deploy on resource-limited hardware. Pruning has enabled efficient compression of large language models (LLMs), yet it is largely understudied in robotics. Surprisingly, we observe that pruning VLA models leads to drastic degradation and increased safety violations. We introduce GLUESTICK, a post-pruning recovery method that restores much of the original model's functionality while retaining sparsity benefits. Our method performs a one-time interpolation between the dense and pruned models in weight-space to compute a corrective term. This correction is used during inference by each pruned layer to recover lost capabilities with minimal overhead. GLUESTICK requires no additional training, is agnostic to the pruning algorithm, and introduces a single hyperparameter that controls the tradeoff between efficiency and accuracy. Across diverse VLA architectures and tasks in manipulation and navigation, GLUESTICK achieves competitive memory efficiency while substantially recovering success rates and reducing safety violations. Additional material can be found at: https://gluestick-vla.github.io/. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_08464 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Don't Run with Scissors: Pruning Breaks VLA Models but They Can Be Recovered Jabbour, Jason Kim, Dong-Ki Smith, Max Patrikar, Jay Ghosal, Radhika Wang, Youhui Agha, Ali Reddi, Vijay Janapa Omidshafiei, Shayegan Robotics Machine Learning Vision-Language-Action (VLA) models have advanced robotic capabilities but remain challenging to deploy on resource-limited hardware. Pruning has enabled efficient compression of large language models (LLMs), yet it is largely understudied in robotics. Surprisingly, we observe that pruning VLA models leads to drastic degradation and increased safety violations. We introduce GLUESTICK, a post-pruning recovery method that restores much of the original model's functionality while retaining sparsity benefits. Our method performs a one-time interpolation between the dense and pruned models in weight-space to compute a corrective term. This correction is used during inference by each pruned layer to recover lost capabilities with minimal overhead. GLUESTICK requires no additional training, is agnostic to the pruning algorithm, and introduces a single hyperparameter that controls the tradeoff between efficiency and accuracy. Across diverse VLA architectures and tasks in manipulation and navigation, GLUESTICK achieves competitive memory efficiency while substantially recovering success rates and reducing safety violations. Additional material can be found at: https://gluestick-vla.github.io/. |
| title | Don't Run with Scissors: Pruning Breaks VLA Models but They Can Be Recovered |
| topic | Robotics Machine Learning |
| url | https://arxiv.org/abs/2510.08464 |