torchrl.record.TensorDictRecorder(out_file_base: str, skip_reset: bool = True, skip: int = 4, in_keys: Optional[Sequence[str]] = None) None[source]

TensorDict recorder.

When the ‘dump’ method is called, this class will save a stack of the tensordict resulting from env.step(td) in a file with a prefix defined by the out_file_base argument.

  • out_file_base (str) – a string defining the prefix of the file where the tensordict will be written.

  • skip_reset (bool) – if True, the first TensorDict of the list will be discarded (usually the tensordict resulting from the call to env.reset()) default: True

  • skip (int) – frame interval for the saved tensordict. default: 4


Access comprehensive developer documentation for PyTorch

View Docs


Get in-depth tutorials for beginners and advanced developers

View Tutorials


Find development resources and get your questions answered

View Resources