TypeError when load_dataset on nvcr.io/nvidia/nemo:25.11

[pcompieta]

Describe the bug

Inside the nvcr.io/nvidia/nemo:25.11 container image, load_dataset("rajpurkar/squad") is failing with below error:

TypeError: must be called with a dataclass type or instance

Steps/Code to reproduce bug

1. Shell into the container docker run --rm -it nvcr.io/nvidia/nemo:25.11
2. load dataset python -c 'from datasets import load_dataset; load_dataset("rajpurkar/squad")'

Expected behavior

Have the dataset downloaded into the proper folder.

Environment overview (please complete the following information)

• Environment location: Docker
• Method of install: apt-get
• If method of install is [Docker], provide docker pull & docker run commands used: see above "repro steps"

Additional context

Docker version 29.1.4

NVIDIA H100
NVIDIA-SMI 590.48.01
Driver Version: 590.48.01
CUDA Version: 13.1

Complete error attached.

load_dataset_error.txt

[akoumpa]

Hi, please note our LLM/VLM work has migrated to megatron bridge repo and nemo automodel

This looks like an issue with datasets' package version, can you try mbridge/automodel instead?

Thank you