> For the complete documentation index, see [llms.txt](https://anonymous-humanoid.gitbook.io/docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://anonymous-humanoid.gitbook.io/docs/getting-started-software/training-environment.md).

# Training Environment

## System and Dependencies 

* Ubuntu 22.04 
* NVIDIA GPU with driver version 535 
* Conda Isaac Lab 2.3.2 
* Isaac Sim 5.1.0 
* Python 3.11

We use Isaac Lab 2.3.2 for policy training. Follow the pip-based installation procedure below. You can refer to our [conda installation tutorial](/docs/getting-started-software/embedded-environment-setup.md#anaconda-miniconda) to install the conda.

{% stepper %}
{% step %}

### Install the NVIDIA Driver

Install the required NVIDIA driver version:

```bash
sudo apt update && sudo apt install nvidia-driver-535
reboot
```

After rebooting, verify the driver installation:

```bash
nvidia-smi
```

{% endstep %}

{% step %}

### Create the Conda Environment

Create and activate a clean Conda environment:

```bash
conda create -n env_isaaclab python=3.11
conda activate env_isaaclab
```

{% endstep %}

{% step %}

### Clone Isaac Lab

Create a projects folder if it is not created yet:

```bash
mkdir ~/projects
```

Navigate to the projects directory and clone Isaac Lab v2.3.2:

```bash
cd ~/projects
git clone --branch v2.3.2 git@github.com:isaac-sim/IsaacLab.git
cd IsaacLab
```

Install the Isaac Sim 5.1.0 and CUDA-enabled versions of PyTorch and Torchvision:

```bash
pip install "isaacsim[all,extscache]==5.1.0" --extra-index-url https://pypi.nvidia.com
pip install -U torch==2.7.0 torchvision==0.22.0 --index-url https://download.pytorch.org/whl/cu128
```

Verify the Isaac Sim Installation

```bash
isaacsim
```

Confirm that Isaac Sim starts correctly before continuing.
{% endstep %}

{% step %}

### Resolve the Optional `CXXABI` Error

On Ubuntu 22.04, Isaac Sim may load the system `libstdc++.so.6`, which provides symbols only up to `CXXABI_1.3.13`. However, packages installed in the Conda environment—such as `icu 78.3`—may require `CXXABI_1.3.15`.

This can produce repeated errors similar to:

```bash
ImportError: /lib/x86_64-linux-gnu/libstdc++.so.6:version `CXXABI_1.3.15' not found
```

When `env_isaaclab` is active, configure Conda to prioritize the environment’s newer `libstdc++` library:

```bash
mkdir -p \
    "$CONDA_PREFIX/etc/conda/activate.d" \
    "$CONDA_PREFIX/etc/conda/deactivate.d"

printf '%s\n' \
    '#!/bin/sh' \
    'export _ISAACLAB_OLD_LD_LIBRARY_PATH="${LD_LIBRARY_PATH-}"' \
    'export LD_LIBRARY_PATH="$CONDA_PREFIX/lib${LD_LIBRARY_PATH:+:$LD_LIBRARY_PATH}"' \
    > "$CONDA_PREFIX/etc/conda/activate.d/zz-libstdcxx.sh"

printf '%s\n' \
    '#!/bin/sh' \
    'export LD_LIBRARY_PATH="${_ISAACLAB_OLD_LD_LIBRARY_PATH-}"' \
    'unset _ISAACLAB_OLD_LD_LIBRARY_PATH' \
    > "$CONDA_PREFIX/etc/conda/deactivate.d/zz-libstdcxx.sh"
```

Reactivate the environment to apply the change:

```bash
conda deactivate
conda activate env_isaaclab
```

This workaround is optional and should be applied only if the `CXXABI_1.3.15` error occurs.
{% endstep %}

{% step %}

### Install Isaac Lab Extensions

From the Isaac Lab repository root, run:

```bash
./isaaclab.sh --install
```

> The executable is `isaaclab.sh`, not `IsaacLab`.
> {% endstep %}

{% step %}

### Clone humanoid-training project

Navigate back to `projects` folder and clone the `humanoid-training` project:

```bash
cd ~/projects
git clone git@github.com:/humanoid-training.git
```

Follow the `README.md` under humanoid folder to train the policy.
{% endstep %}
{% endstepper %}


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://anonymous-humanoid.gitbook.io/docs/getting-started-software/training-environment.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.