Robosuite_task_zoo missing when generating data with NanoLLM container for OpenVLA

[wjdwldud4190]

robosuite_task_zoo_log.txt (8.4 KB)

Hello,

I’m trying to generate data using the nano_llm container on my Jetson AGX Orin (JetPack 6.2.1, L4T 36.4.4, CUDA 12.6). However, I encounter the following warnings and error during startup:

OpenVLA - NVIDIA Jetson AI Lab

jetson-containers run $(autotag nano_llm) \
  python3 -m mimicgen.generate \
      --tasks Stack_D4 \
      --episodes 100 \
      --output /data/datasets/mimicgen \
      --cameras agentview \
      --camera-width 224 \
      --camera-height 224


....
[robosuite WARNING] No private macro file found! (macros.py:53)
[robosuite WARNING] It is recommended to use a private macro file (macros.py:54)
[robosuite WARNING] To setup, run: python /opt/robosuite/robosuite/scripts/setup_macros.py (macros.py:55)

WARNING: robosuite task zoo environments not imported, possibly because robosuite_task_zoo is not installed...

When I tried installing robosuite_task_zoo manually inside the container, pip failed to fetch it from the Jetson AI Lab custom index:

root@ubuntu:/# pip install robosuite_task_zoo
Looking in indexes: http://jetson.webredirect.org/jp6/cu126, https://pypi.ngc.nvidia.com
WARNING: Retrying (Retry(total=4, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ConnectTimeoutError(<pip._vendor.urllib3.connection.HTTPSConnection object at 0xffffbbf8f250>, 'Connection to pypi.jetson-ai-lab.dev timed out. (connect timeout=100.0)')': /jp6/cu126/robosuite-task-zoo/

So currently the package can’t be installed from inside the container. I attached my log.

I also found a related GitHub issue reported here:
🔗 https://github.com/dusty-nv/jetson-containers/issues/741
…but it seems no solution has been provided yet.

[carolyuu]

*** Please note that this reply is generated by LLM automatically ***

Based on the provided context, it appears that you are experiencing issues with installing a package inside a container on your Jetson AGX Orin device, which is running JetPack 6.2.1 (L4T 36.4.4) and CUDA 12.6. The package you are trying to install is robosuite-task-zoo-log-txt, which is not available on the Jetson AI Lab custom index.

The warnings and error you are encountering during startup are likely related to the fact that the package cannot be installed from inside the container. The log you attached suggests that pip is failing to fetch the package from the Jetson AI Lab custom index.

Comparing your issue to the related GitHub issue you mentioned (https://github.com/dustynv/jetson-containers/issues/741), it seems that the issue is still unresolved. However, there are a few potential solutions you could try:

1. Use the --index-url flag: You can try installing the package using the --index-url flag, which allows you to specify a custom index URL. For example:

pip install --index-url https://pypi.jetson-ai-lab.dev/sbsa/cu128 robosuite-task-zoo-log-txt

This may allow you to install the package from the Jetson AI Lab custom index.

2. Use a different package manager: You can try using a different package manager, such as apt, to install the package. For example:

apt install robosuite-task-zoo-log-txt

This may allow you to install the package from the Ubuntu repositories.

3. Build the package from source: If the package is not available on the Jetson AI Lab custom index or the Ubuntu repositories, you may need to build it from source. This can be a more complex process, but it may allow you to install the package.

4. Contact the package maintainer: If none of the above solutions work, you may want to contact the package maintainer to see if they can provide a solution or update the package to make it compatible with your Jetson AGX Orin device.

In summary, while there is no straightforward solution to your issue, you can try using the --index-url flag, a different package manager, building the package from source, or contacting the package maintainer to see if they can provide a solution.

*** Please give the thumbs up if you get answers you like. Or provide feedback to help us improve the answer. ***

[wjdwldud4190]

I tried the suggestions to install from the Jetson AI Lab Python index and via apt. I’m still blocked.

1. Use the --index-url flag:

root@ubuntu:/# pip install --index-url https://pypi.jetson-ai-lab.dev/sbsa/cu128 robosuite-task-zoo-log-txt
Looking in indexes: https://pypi.jetson-ai-lab.dev/sbsa/cu128, https://pypi.ngc.nvidia.com
WARNING: Retrying (Retry(total=4, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ConnectTimeoutError(<pip._vendor.urllib3.connection.HTTPSConnection object at 0xffffbce8ee00>, 'Connection to pypi.jetson-ai-lab.dev timed out. (connect timeout=100.0)')': /sbsa/cu128/robosuite-task-zoo-log-txt/
WARNING: Retrying (Retry(total=3, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ConnectTimeoutError(<pip._vendor.urllib3.connection.HTTPSConnection object at 0xffffbce8f100>, 'Connection to pypi.jetson-ai-lab.dev timed out. (connect timeout=100.0)')': /sbsa/cu128/robosuite-task-zoo-log-txt/

2. Use a different package manage

root@ubuntu:/# apt install robosuite-task-zoo-log-txt
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
E: Unable to locate package robosuite-task-zoo-log-txt

[AastaLLL]

Hi,

Please try below to install the robosuite_task_zoo manually.

# git clone https://github.com/ARISE-Initiative/robosuite-task-zoo.git
# cd robosuite-task-zoo/
# pip install -e . --isolated

Then you might meet another issue related to missing Google Drive files.

    assert url_is_alive(url), "@download_url_from_gdrive got unreachable url: {}".format(url)

Then please check the link below to update it in the HuggingFace:

No mimicgen data

Hi, The data has been moved from google drive to the huggingface. Thanks.

Thanks.

[wjdwldud4190]

hi, thanks for reply.

I recently did a fresh git pull from:

origin  https://github.com/dusty-nv/jetson-containers (fetch)
origin  https://github.com/dusty-nv/jetson-containers (push)

Then I re-ran the following command:

python3 -m mimicgen.generate \
    --tasks Stack_D4 \
    --episodes 100 \
    --output /data/datasets/mimicgen \
    --cameras agentview \
    --camera-width 224 \
    --camera-height 224

However, after this git pull, I now get a different error than before.

and I installed the robosuite_task_zoo using your comment.

I tested with nanollm version r36.3 and r36.4, but they show the same issue.

mimicgeneration_log.txt (12.0 KB)

thanks

[AastaLLL]

Hi,

We test this with dustynv/nano_llm:r36.4.0 and it can work correctly.
Based on your error below:

Traceback (most recent call last):
  File "/opt/robosuite/robosuite/renderers/context/egl_context.py", line 155, in __del__
    self.free()
  File "/opt/robosuite/robosuite/renderers/context/egl_context.py", line 149, in free
    EGL.eglMakeCurrent(EGL_DISPLAY, EGL.EGL_NO_SURFACE, EGL.EGL_NO_SURFACE, EGL.EGL_NO_CONTEXT)
  File "/usr/local/lib/python3.10/dist-packages/OpenGL/error.py", line 230, in glCheckError
    raise self._errorClass(
OpenGL.raw.EGL._errors.EGLError: EGLError(
        err = EGL_NOT_INITIALIZED,
        baseOperation = eglMakeCurrent,
        cArguments = (
                <OpenGL._opaque.EGLDisplay_pointer object at 0xfffed11d0c40>,
                <OpenGL._opaque.EGLSurface_pointer object at 0xffff17ac2bc0>,
                <OpenGL._opaque.EGLSurface_pointer object at 0xffff17ac2bc0>,
                <OpenGL._opaque.EGLContext_pointer object at 0xffff17ac2cc0>,
        ),
        result = 0
)

Could you try to export the DISPLAY before launching the container?

$ export DISPLAY=:0
$ xhost + 
$ jetson-containers run $(autotag nano_llm)
...

Thanks.