Get no error while using list object in bg.set_batch_num_nodes() but will get an AttributeError when using AvgPool to pool the batched graph

[NormanZyq]

🐛 Bug

If use a list object as the argument of set_batch_num_nodes(), we get no error, and it seems the graph is successfully batched with the number of nodes. However, if I use AvgPooling to pool the batched graph, I would get an error:

AttributeError: 'list' object has no attribute 'dtype'

It's obvious that the list is not supported in this case.
If I use tensor, everything works fine, as the doc mentioned.

To Reproduce

Steps to reproduce the behavior:

Please refer to the notebook:
https://colab.research.google.com/drive/1n-tfuPp3R59olfxgN4I9NJ95bNUmdsdJ?usp=sharing

Expected behavior

If I use a list, there should be an error raised, or it can be transformed into tensor automatically

Environment

• DGL Version (e.g., 1.0): 1.1.2
• Backend Library & Version (e.g., PyTorch 0.4.1, MXNet/Gluon 1.3): 2.1.0+cu118
• OS (e.g., Linux): Linux
• How you installed DGL (conda, pip, source): pip
• Build command you used (if compiling from source):
• Python version: 3.10.12
• CUDA/cuDNN version (if applicable):
• GPU models and configuration (e.g. V100):
• Any other relevant information:

Additional context

[Rhett-Ying]

Thanks for reporting this. We do need to avoid passing unexpected data.