Repository Issue Activity (beta)

#991 Numpy > 2 Support for openpi-client?

#480 Jax Compilation Takes forever

#988 (video) Wrist cam image does not attend to target [PI05 inference, SO-101 arm]

#708 Question about the fast tokenizer in pi05

#989 Performance gap between PyTorch and JAX fine-tuning of pi0.5_base on LIBERO / LIBERO-Plus OOD

#987 In the setup function inside the Module class in gemma.py, should the static_argnums parameter in nn.remat be set to (5,) or (6,)?

#986 convert_aloha_data_to_lerobot.py is outdated and extremely memory-heavy

#985 How to properly merge LoRA weights back into a Pi0.5 (pi05) base checkpoint?

#983 The inference timing script is incorrect

#982 How to export TRT from Pytorch of the Pi0.5?

#980 Will new models be open-sourced, and will dependencies be updated to support newer hardware (e.g., CUDA 13.x, NVIDIA DGX Spark)?

#974 Do you have a TrainConfig for fine-tuning pi0.5 libero lora? Is this configuration correct?

#973 Pi0.5 reproduction

#706 dataset format versioning

#849 Pi0 fast libero checkpoint no longer works with new repo?

#714 dimension of action_in_proj and action_out_proj in pi0_pytorch.py should be fixed

#961 Does high-level policy model and low-level policy model are same model in MEM and pi0.7?

#912 [pi0.5] Robot reaches object but fails to grasp, keeps pushing object with small actions

#967 Can I apply incremental learning here? What suggestions do you have?

#842 [Feature Proposal] Add LoRA training support for pi0.5 (PyTorch backend)

#947 SigLIP learned posemb has extreme L2 norm disparity across the patch grid, when loading openpi's official pi05 ckpt

#963 111

#810 Pi05 output mismatch between PyTorch and Jax weights

#959 Questions about the details of pi05 model training and deployment

#958 examples/convert_jax_model_to_pytorch.py silently drops LoRA adapter weights for LoRA-finetuned checkpoints