v0.3.5

Changelog

• Support hashing the folded_tensor.length field (via a UserList), which is convenient for caching
• Improve error messaging when refolding with missing dims

What's Changed

• Drop codecov by @percevalw in #10
• feat: hashable lengths by @percevalw in #12
• Improve missing dim error by @percevalw in #13

Full Changelog: v0.3.4...v0.3.5

v0.3.4

What's Changed

• Fix a data_dims access issue
• Marginally improve the speed of handling FoldedTensors in standard torch operations
• Use default torch types (e.g. torch.float32 or torch.torch64)

Pull Requests

• Speedups & fixes by @percevalw in #9

Full Changelog: v0.3.3...v0.3.4

v0.3.3

What's changed

• Handle empty inputs (e.g. as_folded_tensor([[[], []], [[]]])) by returning an empty tensor
• Correctly bubble errors when converting inputs with varying deepness (e.g. as_folded_tensor([1, [2, 3]]))

Pull Requests

• Handle empty inputs and show better errors for inputs with varying deepness by @percevalw in #7
• Bump version to 0.3.3 by @percevalw in #8

Full Changelog: v0.3.2...v0.3.3

v0.3.2

Changelog

• Allow to use as_folded_tensor with no args, as a simple padding function

What's Changed

• Empty args by @percevalw in #5

Full Changelog: v0.3.1...v0.3.2

v0.3.1

What's changed

• Enable sharing FoldedTensor instances in a multiprocessing + cuda context by autocloning the indexer before fork-pickling an instance
• Distribute arm64 wheels for macOS

Pull Requests

• Enable sharing in multiprocessing + cuda context by @percevalw in #3
• Build for macos arm arch by @percevalw in #4

Full Changelog: v0.3.0...v0.3.1

v0.3.0

What's Changed

• Allow dims after last foldable dim during list conversion (e.g. embeddings)

Full Changelog: v0.2.2...v0.3.0

v0.2.2

GitHub release !

• Support for arbitrary numbers of nested dimensions
• No computational overhead when dealing with already padded tensors
• Dynamic re-padding (or refolding) of data based on stored inner lengths
• Automatic mask generation and updating whenever the tensor is refolded
• C++ optimized code for fast data loading from Python lists and refolding
• Flexibility in data representation, making it easy to switch between different layouts when needed
• No pytorch binary dependency (allowing us to distribute prebuilt binaries)

Full Changelog: https://github.com/aphp/foldedtensor/commits/v0.2.2