How can I change the "batch_size"

[dengzx7]

I directly change the "batch_size" in file "execute_cora.py" from 1 to 10, and the following error occurred.

ValueError: Cannot feed value of shape (1, 2708, 1433) for Tensor 'input/Placeholder:0', which has shape '(10, 2708, 1433)'

So how can I change the "batch_size" when training or running the model ? Thx

[liujian19911023]

I directly change the "batch_size" in file "execute_cora.py" from 1 to 10, and the following error occurred.

ValueError: Cannot feed value of shape (1, 2708, 1433) for Tensor 'input/Placeholder:0', which has shape '(10, 2708, 1433)'

So how can I change the "batch_size" when training or running the model ? Thx

change placeholder batch_size to None，but this is not useful，every epoch the model use full data， not node level。when graph is very big， need the change code to support node level train。

[tanjia123456]

hello, do you know how to modify the value of the batch-size? It seems can't change it in the layer.py If you don't mind ,could you please share experience of it?

[tanjia123456]

hello, do you know how to modify the value of the batch-size? It seems can't change it in the layer.py If you don't mind ,could you please share experience of it? This problem has been bothering me for a week, and I really can’t find a clue

[wenwen19980927]

hi， do you know how to modify the batchsize now? ^^

[hazdzz]

GAT is a full batch GNN model. If you require a mini-batch model, you need to graph sampling first.