pytorch_pfn_extras#

Functions

`pytorch_pfn_extras.as_ndarray`(tensor)	Creates a numpy.ndarray or cupy.ndarray from torch.Tensor.
`pytorch_pfn_extras.as_numpy_dtype`(torch_dtype)	Returns NumPy dtype for the given PyTorch dtype.
`pytorch_pfn_extras.compile`(module[, ...])	Compiles a module and an optimizer in a single graph using the provided backend.
`pytorch_pfn_extras.from_ndarray`(ndarray)	Creates a torch.Tensor from a numpy.ndarray or cupy.ndarray.
`pytorch_pfn_extras.from_numpy_dtype`(numpy_dtype)	Returns PyTorch dtype for the given NumPy dtype.
`pytorch_pfn_extras.get_xp`(obj)	Returns a module of ndarray implementation (numpy or cupy) for the given obj.
`pytorch_pfn_extras.map`(func, iterable[, ...])
`pytorch_pfn_extras.requires`(version[, package])
`pytorch_pfn_extras.to`(module_or_tensor, ...)	A function to transfer the given object to the given device.

Modules

`pytorch_pfn_extras.config`
`pytorch_pfn_extras.config_types`
`pytorch_pfn_extras.cuda`
`pytorch_pfn_extras.dataloaders`
`pytorch_pfn_extras.dataset`
`pytorch_pfn_extras.distributed`
`pytorch_pfn_extras.engine`
`pytorch_pfn_extras.handler`
`pytorch_pfn_extras.logging`
`pytorch_pfn_extras.nn`
`pytorch_pfn_extras.onnx`
`pytorch_pfn_extras.ops`
`pytorch_pfn_extras.profiler`
`pytorch_pfn_extras.reporting`
`pytorch_pfn_extras.runtime`
`pytorch_pfn_extras.testing`
`pytorch_pfn_extras.torchscript`
`pytorch_pfn_extras.training`
`pytorch_pfn_extras.utils`
`pytorch_pfn_extras.writing`