openpi
openpi copied to clipboard
When I use multi-GPU training, the following error occurs:“Traceback (most recent call last): File "/liury/src/openpi/scripts/train.py", line 304, in main(_config.cli()) File "/liury/src/openpi/scripts/train.py", line 285, in main train_state, info = ptrain_step(train_rng, train_state,...
I noticed that pi0 performs well on Libero Spatial, Libero Object, Libero Goal, and Libero 10 datasets. However, the model I tried to pretrain on libero 90 does not work...
Hello, thanks for your great work first! I recently downloaded the LIBERO dataset, and I noticed something confusing regarding the representation of state and action in the feature files. -...
Cell In[5], line 7 2 config = dataclasses.replace(config, batch_size=2) 4 # Load a single batch of data. This is the same data that will be used during training. 5 #...
Environment: Ubuntu 24.04, RTX5090, CUDA 12.6.3 When I run XLA_PYTHON_CLIENT_MEM_FRACTION=0.9 uv run scripts/train.py pi0_fast_libero --exp-name=my_experiment --overwrite I get the following error: XLA_PYTHON_CLIENT_MEM_FRACTION=0.9 uv run scripts/train.py pi0_fast_libero --exp-name=my_experiment --overwrite warning: The...
hi, I see that executing `uv run scripts/compute_norm_stats.py` is a must when fine-tuning on libero dataset. however, when i fine-tune on droid dataset, and if i skip the compute norm...
I'm trying to deploy the OpenPI inference server by following the official instructions. The server seems to start correctly, but when I run the client, it fails with the following...
python 3.13 install openpi will meet problem , bazel cannot find. But use python 3.11 can install smoothly.
Hi, I'm trying to fine-tune PI0 base model on Droid dataset. I noticed there is only a PI0-Fast-Finetune version of TrainConfig, so I defined a PI0-Finetune version TrainConfig as follows:...