17 lines
901 B
Docker
17 lines
901 B
Docker
# we will run `docker run -it --rm -v ./dataset_dir:/dataset YOUR_IMAGE` to run the downloading container.
|
|
# The provider will mount their `dataset_dir` to `/dataset` directory in the container.
|
|
# So, you should save your dataset into `/dataset` so that our `training container` can access your dataset.
|
|
|
|
FROM ubuntu:22.04
|
|
|
|
# This is our example script:
|
|
# First, we create the directory named `/dataset` so that it can link to the provider's host.
|
|
# Then, we use apt to download `wget` so that we can use wget to download the dataset from our target URL.
|
|
RUN mkdir /dataset && apt-get update && apt-get install -y wget
|
|
|
|
# And we get the dataset file from the target URL and save it into /dataset
|
|
# Then, we unzip it so that the `training container` can access them directly.
|
|
WORKDIR /dataset
|
|
CMD pwd && wget https://www.cs.toronto.edu/\~kriz/cifar-10-python.tar.gz && tar zxvf cifar-10-python.tar.gz
|
|
|