Downloads · 30 days
120
21% of all-time downloads
dynamic-vla-DOM
hzxie/dynamic-vla-DOM
dynamic-vla-DOM is a robotics model from hzxie. 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 lerobot. The card lists the license as other.
DynamicVLA is a vision-language-action model for dynamic object manipulation. It is designed to handle dynamic scenes that require fast perception, temporal anticipation, and continuous control.
All-time downloads
575
Public
Parameters
430M
1.1 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors1.1 GB · 100%
Parameter types
How the weights are stored.
BF16301M · 70%
Base models
Datasets
Model card
From the Hugging Face model README
Model Card for DynamicVLA
DynamicVLA is a vision-language-action model for dynamic object manipulation. It is designed to handle dynamic scenes that require fast perception, temporal anticipation, and continuous control.
This model is trained and evaluated using the official DynamicVLA codebase. For full setup, training, and benchmarking instructions, please refer to the repository README.
How to Get Started with the Model
For a complete walkthrough, see the official DynamicVLA repository. Below is the short version for training and running inference/evaluation.
Train from scratch
From the PROJECT_ROOT/dynamic-vla directory, run:
torchrun --nnodes=1 --nproc_per_node=8 --standalone run.py \
-c configs/dynamicvla.yaml \
-d hzxie/DOM
Evaluate the policy / run inference
# 1. start evaluation server
python3 simulations/evaluate.py \
--scene_dir ../scenes \
--output_dir ../output/evaluation \
--env_cfg ../test-envs.txt \
--enable_cameras --headless -n 20 --save
# 2. run policy inference
python3 scripts/inference.py \
-p /path/to/vla-checkpoint \
-r euler -d -s
- Task
- Robotics
- Library
- lerobot
- License
- other
- Created
- Feb 4, 2026
- Updated
- Apr 23, 2026