ALOHA机器人 ACT算法复现

ALOHA机器人 ACT算法复现（在4自由度小玩具上的复现）

---

原论文标题 A Low-cost Open-source Hardware System for Bimanual Teleoperation， 作者单位：斯坦福大学。实现了一种端到端的模仿学习算法。

安装配置

下载

首先下载项目

git clone https://github.com/agilexrobotics/act-plus-plus.git

配置环境

配置环境，原工程没有撰写虚拟环境配置，以下给出

torch==1.11.0+cu113
torchvision==0.12.0+cu113 
torchaudio==0.11.0
pyquaternion
pyyaml
rospkg
pexpect
mujoco==2.3.7
dm_control==1.0.14
opencv-python
matplotlib
einops
packaging
h5py
ipython
diffusers
wandb

安装软件需求

cd detr && pip install -v -e .

修改数据存储路径，在constant.py文件中:

DATA_DIR = './data'

错误处理

---

• 报错1

For debugging consider passing CUDA_LAUNCH_BLOCKING=1.
Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.

• 解决方案：

export CUDA_LAUNCH_BLOCKING=1
# 或者
pip uninstall nvidia-cublas-*

---

• 报错2

Segmentation fault (core dumped)

• 解决方案：
  检查pytorch版本。或者直接运行

pip install torch==1.11.0+cu113 torchvision==0.12.0+cu113 torchaudio==0.11.0 --extra-index-url https://download.pytorch.org/whl/cu113

• 报错3

NotImplementedError: There were no tensor arguments to this function (e.g., you passed an empty list of Tensors), but no fallback function is registered for schema aten::_cat.  This usually means that this function requires a non-empty list of Tensors, or that you (the operator writer) forgot to register a fallback function.  Available functions are [CPU, CUDA, QuantizedCPU, BackendSelect, Python, Named, Conjugate, Negative, ZeroTensor, ADInplaceOrView, AutogradOther, AutogradCPU, AutogradCUDA, AutogradXLA, AutogradLazy, AutogradXPU, AutogradMLC, AutogradHPU, AutogradNestedTensor, AutogradPrivateUse1, AutogradPrivateUse2, AutogradPrivateUse3, Tracer, AutocastCPU, Autocast, Batched, VmapMode, Functionalize].

检查数据的路径是否和constant.py保持一致。检查task name参数指定的数据是否存在。

训练

python3 imitate_episodes.py --task_name sim_transfer_cube_scripted --ckpt_dir trainings --policy_class ACT --kl_weight 1 --chunk_size 10 --hidden_dim 512 --batch_size 1 --dim_feedforward 3200  --lr 1e-5 --seed 0 --num_steps 2000

修改

未完待续。。。