Not enought space to install xgboost on Retool Workflow

💬 Workflows
1

Hi everyone!!

I'd like to test a workflow where I use an existing xgboost model to predict some values based on data input from the app and then return it to the app.

So far I'm simply stuck at installing the xgboost library, it seems there's not enough space, see log:

Collecting msgspec==0.13.1
Downloading msgspec-0.13.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (177 kB)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 177.8/177.8 kB 6.0 MB/s eta 0:00:00
Collecting mypy==1.5.1
Downloading mypy-1.5.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (12.1 MB)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 12.1/12.1 MB 197.3 MB/s eta 0:00:00
Collecting pip==25.0
Downloading pip-25.0-py3-none-any.whl (1.8 MB)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 1.8/1.8 MB 284.6 MB/s eta 0:00:00
Collecting xgboost==2.1.4
Downloading xgboost-2.1.4-py3-none-manylinux_2_28_x86_64.whl (223.6 MB)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 223.6/223.6 MB 36.7 MB/s eta 0:00:00
Collecting tomli>=1.1.0
Downloading tomli-2.2.1-py3-none-any.whl (14 kB)
Collecting mypy-extensions>=1.0.0
Downloading mypy_extensions-1.0.0-py3-none-any.whl (4.7 kB)
Collecting typing-extensions>=4.1.0
Downloading typing_extensions-4.12.2-py3-none-any.whl (37 kB)
Collecting numpy
Downloading numpy-2.2.2-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (16.4 MB)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 16.4/16.4 MB 193.5 MB/s eta 0:00:00
Collecting nvidia-nccl-cu12
Downloading nvidia_nccl_cu12-2.25.1-py3-none-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (201.4 MB)
━━━━━━━━━━━━━━━━━╸ 95.2/201.4 MB 352.7 MB/s eta 0:00:01
ERROR: Could not install packages due to an OSError: [Errno 28] No space left on device

Since I'm quite new to Workflows I really don't know if that is normal or not. I could install pip but which seems to be much lighter.

Also it seems to be installing other librairies that I didn't click on (numpy, tomli, etc.).

Any idea on how to solve this?

B.

3

Hi everyone,

A small up here :slight_smile:

Any idea?

B.

4

Hey @Baptiste_LC! Thanks for reaching out.

Assuming you're running a cloud instance, each workflow run is allocated a total of ~1GB RAM storage that is shared among all processes and installations. It looks like the default xgboost library and its dependencies eats up a big chunk of that. You might want to try using the xgboost-cpu version, instead!

5

Just checking in, @Baptiste_LC - have you had a chance to test out the above recommendation?