chauhanjatin10 / graphsfewshot Goto Github PK

hello,
Thank you for your sharing the code.
In section 5.2, you write

The results for all the datasets in various q-shot scenarios

And in your code in src/graphcnn.py

				selected_idx = np.random.permutation(len(train_graphs))[:args.batch_size]
				batch_graph = [train_graphs[idx] for idx in selected_idx]

I want to ask that:
The "shot" in your article is different from "N-way, K-Shot" in Image Task Meta Learning.
In Image Task Meta-Learning, we first pick out N classes from meta-train dataset, whose classes is disjoint with that of meta-test dataset. And then, for each picked out classes, K shots are picked out.
So, ... N * K samples are picked out. And in your article, you just use q shot, "q" has the same meaning with "N * K"?

yours,
@WMF1997

Hi, thanks for your code. Could you explain the content of Letter-high.txt? I have no idea about the content format.