v23.06.00

Update merlin-core version specifier to match current release (#154)

v23.08.00: update workflow on which docs pr preview depends (#141)

Co-authored-by: Oliver Holworthy <oholworthy@nvidia.com>
Co-authored-by: Karl Higley <kmhigley@gmail.com>

v23.05.00

What’s Changed

🐜 Bug Fixes

• Adjust tag propagation in EmbeddingOperator @karlhigley (#146)

🚀 Features

• Use Dictionary lookup for supplied IDs to Embedding Operator @oliverholworthy (#148)
• Fix the output type of Padding op @sararb (#145)

🔧 Maintenance

• Update checks of HAS_GPU in tests to handle case where cudf is not installed @oliverholworthy (#118)
• Update tag pattern in GitHub Workflows @oliverholworthy (#143)
• Add workflows to check base branch and set stable branch @oliverholworthy (#142)
• Remove use of deprecated numpy aliases of builtin types @oliverholworthy (#144)
• Update requirements for Merlin packages to minimum version of 23.04 @oliverholworthy (#140)

v23.04.00

What’s Changed

⚠ Breaking Changes

• Return Dense Tensors for Fixed List Columns @oliverholworthy (#112)
• Remove sparse tensor output type for list features @oliverholworthy (#103)
• Update dataloader to provide new output structure @bschifferer (#101)

🐜 Bug Fixes

• Apply device context when converting to torch tensors @edknv (#135)
• Revert "Set device for torch tensors with gpu > 1 (#132)" @edknv (#134)
• Set device for torch tensors with gpu > 1 @edknv (#132)
• Apply map function when peeking in tensorflow @edknv (#129)
• remove numpy tensorflow conversion logic in tf dataloader @jperez999 (#122)
• Change row lengths unit tests to use cpu and single gpu @edknv (#115)
• Put row lengths on the same device on gpu @edknv (#113)

🚀 Features

• Adding Pad op @jperez999 (#125)
• fixing imports for tf and torch in dataloader @jperez999 (#127)
• Add downstream repo tests @edknv (#126)
• imports for GPU packages from compat @jperez999 (#119)
• adding array version of dataloader @jperez999 (#111)
• Return Dense Tensors for Fixed List Columns @oliverholworthy (#112)
• Remove sparse tensor output type for list features @oliverholworthy (#103)
• Update dataloader to provide new output structure @bschifferer (#101)
• Apply the EMBEDDING tag to the output of the embedding operators @karlhigley (#110)

🔧 Maintenance

• Warn users when a batch size may not work on GPU @edknv (#136)
• Revert "Set device for torch tensors with gpu > 1 (#132)" @edknv (#134)
• Adding Pad op @jperez999 (#125)
• fixing imports for tf and torch in dataloader @jperez999 (#127)
• Add downstream repo tests @edknv (#126)
• Array loaderbase merge @jperez999 (#121)
• imports for GPU packages from compat @jperez999 (#119)
• Remove reset of dataloader in len method of ArrayLoader @oliverholworthy (#123)
• remove numpy tensorflow conversion logic in tf dataloader @jperez999 (#122)
• replace dictarray for tensortable in transforms functionality @jperez999 (#120)
• Remove reshape to speed up loading of scalar features in TensorFlow @oliverholworthy (#116)
• Change row lengths unit tests to use cpu and single gpu @edknv (#115)
• Update default value of cpu to None in dataset fixture @oliverholworthy (#114)
• Correct cpu-packages config @oliverholworthy (#109)

v23.02.01

What's Changed

🔧 Maintenance

• Refactor ragged tensor processing for readbility and improved performance @oliverholworthy (#104)
• Restore support for passing device in CPU-only environment @oliverholworthy (#107)
• Correct cpu-packages config - rename build -> build-packages @oliverholworthy (#109)

Full Changelog: v23.02.00...v23.02.01

v23.02.00

What’s Changed

🚀 Features

• add map to torch Loader @radekosmulski (#90)

📄 Documentation

• add example for running on multiple GPUs using Horovod @radekosmulski (#81)

🔧 Maintenance

• Replace nnzs with row_lengths for clarity @karlhigley (#99)
• fix gpu visibility on docker @jperez999 (#98)
• Add Jenkinsfile @AyodeAwe (#49)
• Split CPU tests and building packages into separate Github Actions @karlhigley (#93)
• Check Input Schema setter with list feature @oliverholworthy (#87)
• Update pytorch getting started test and setup examples tests to run in CI @oliverholworthy (#63)
• activate gha runner for gpu-ci @jperez999 (#88)

New Contributors

• @karlhigley made their first contribution in #93
• @AyodeAwe made their first contribution in #49

Full Changelog: v0.0.4...v23.02.00

v0.0.4

Use tf.function for list column operations (#89)

v0.0.3

What’s Changed

🐜 Bug Fixes

• Put values and offsets on the same device on gpu(#45)

🚀 Features

• Dl transforms (#41)

📄 Documentation

• docs: Add basic SEO configuration (#43)

🔧 Maintenance

• add capability to target branch for unit test in cpu on tox (#58)
• Add lint workflow using pre-commit (#55)
• clear init to ensure loading of framework specific ops only (#48)
• add tox ini as entrypoint for tests in cpu gha (#42)

v0.0.2

Add MANIFEST.in (#40)

v0.0.1

Update README.md