search

tristandeleu / pytorch-meta

A collection of extensions and data-loaders for few-shot learning & meta-learning in PyTorch
https://tristandeleu.github.io/pytorch-meta/
MIT License
1.97k stars 256 forks source link

Questions about downloading datasets #133

Closed liguge closed 2 years ago

liguge commented 3 years ago

I am a novice, and I have only been exposed to this subject not long . Your code is of great reference for my understanding of meta-learning. When I run the code, the following error is displayed. After searching for a long time, I haven't found how to solve it. I want to ask for advice. thanks

sage: Model-Agnostic Meta-Learning (MAML) [-h] [--num-shots NUM_SHOTS] [--num-ways NUM_WAYS] [--first-order] [--step-size STEP_SIZE] [--hidden-size HIDDEN_SIZE] [--output-folder OUTPUT_FOLDER] [--batch-size BATCH_SIZE] [--num-batches NUM_BATCHES] [--num-workers NUM_WORKERS] [--download] [--use-cuda] folder Model-Agnostic Meta-Learning (MAML): error: the following arguments are required: folder

tristandeleu commented 3 years ago

I'm sorry for the late reply. The folder argument is the path to the folder where the dataset (Omniglot here) is. For example, you can run it as

python maml.py data --num-shots 5 --num-ways 5 --download

That will download the dataset into the data/omniglot folder.