Shortcuts

OpenMLEnv

torchrl.envs.libs.openml.OpenMLEnv(*args, _inplace_update=False, _batch_locked=True, **kwargs)[source]

An environment interface to OpenML data to be used in bandits contexts.

Parameters:
  • dataset_name (str) – the following datasets are supported: "adult_num", "adult_onehot", "mushroom_num", "mushroom_onehot", "covertype", "shuttle" and "magic".

  • device (torch.device or compatible, optional) – the device where the input and output data is to be expected. Defaults to "cpu".

  • batch_size (torch.Size or compatible, optional) – the batch size of the environment, ie. the number of elements samples and returned when a reset() is called. Defaults to an empty batch size, ie. one element is sampled at a time.

Examples

>>> env = OpenMLEnv("adult_onehot", batch_size=[2, 3])
>>> print(env.reset())
TensorDict(
    fields={
        done: Tensor(shape=torch.Size([2, 3, 1]), device=cpu, dtype=torch.bool, is_shared=False),
        observation: Tensor(shape=torch.Size([2, 3, 106]), device=cpu, dtype=torch.float32, is_shared=False),
        reward: Tensor(shape=torch.Size([2, 3, 1]), device=cpu, dtype=torch.float32, is_shared=False),
        y: Tensor(shape=torch.Size([2, 3]), device=cpu, dtype=torch.int64, is_shared=False)},
    batch_size=torch.Size([2, 3]),
    device=cpu,
    is_shared=False)

Docs

Access comprehensive developer documentation for PyTorch

View Docs

Tutorials

Get in-depth tutorials for beginners and advanced developers

View Tutorials

Resources

Find development resources and get your questions answered

View Resources