OpenVLA 7B Fine Tuned on LIBERO Object This model was produced by fine tuning the OpenVLA 7B model via LoRA (r=32) on the LIBERO Object dataset from the LIBERO simulation benchmark. We made a few modifications to the training dataset to improve final performance (see the OpenVLA paper for details). Below are the hyperparameters we used for all LIBERO experiments: Hardware: 8 x A100 GPUs with 80GB memory Fine tuned with LoRA: use lora == True , lora rank == 32 , lora dropout == 0.0 Learning rate: 5e 4 Batch size: 128 (8 GPUs x 16 samples each) Number of training gradient steps: 50K No quantization at train or test time No gradient accumulation (i.e. grad accumulation steps == 1 ) shuffle buffer size == 100 000 Image augmentations: Random crop, color jitter (see training code for details) Usage Instructions See the OpenVLA GitHub README for instructions on how to run and evaluate this model in the LIBERO simulator. Citation BibTeX:
We use cookies for essential functionality and analytics. You can accept or reject analytics cookies.Cookie policy