Lightning-AI/lightning-thunder 0.2.4

0.2.4

Repository: Lightning-AI/lightning-thunder

Tag: 0.2.4

Published: 2025-06-24T10:49:45Z

Prerelease: no

Release notes:

What's Changed

• cleaning skipif for past Torch dev versions by @Borda in https://github.com/Lightning-AI/lightning-thunder/pull/2125
• fix missing images when released on PyPI by @Borda in https://github.com/Lightning-AI/lightning-thunder/pull/2130
• Add custom decompositions for cross entropy loss for the nvfuser executor by @protonu in https://github.com/Lightning-AI/lightning-thunder/pull/2043
• [pre-commit.ci] pre-commit suggestions by @pre-commit-ci in https://github.com/Lightning-AI/lightning-thunder/pull/2139
• Remove qualified access to methods from autodiff by @riccardofelluga in https://github.com/Lightning-AI/lightning-thunder/pull/2147
• add non-None check for torch.utils.collect_env.get_pip_packages outputs by @crcrpar in https://github.com/Lightning-AI/lightning-thunder/pull/2124
• Print repro command when test_core_vs_torch_consistency fails with sample index specified by @crcrpar in https://github.com/Lightning-AI/lightning-thunder/pull/2131
• Set input_quantizer.internal to True by @beverlylytle in https://github.com/Lightning-AI/lightning-thunder/pull/2146
• Bump transformers from 4.50.3 to 4.52.4 by @dependabot in https://github.com/Lightning-AI/lightning-thunder/pull/2160
• Update ipython[all] requirement from ~=8.36.0 to ~=8.37.0 by @dependabot in https://github.com/Lightning-AI/lightning-thunder/pull/2159
• Update coverage requirement from ~=7.6.8 to ~=7.8.2 by @dependabot in https://github.com/Lightning-AI/lightning-thunder/pull/2158
• TE: update test to be more stable by @kshitij12345 in https://github.com/Lightning-AI/lightning-thunder/pull/2156
• Bump pytest-timeout from 2.3.1 to 2.4.0 by @dependabot in https://github.com/Lightning-AI/lightning-thunder/pull/2161
• Update dependabot - reviewers by @Borda in https://github.com/Lightning-AI/lightning-thunder/pull/2162
• Fix autodiff joint trace dataflow and in-place ops in higher order functions by @riccardofelluga in https://github.com/Lightning-AI/lightning-thunder/pull/2143
• Reduces the test time by @kiya00 in https://github.com/Lightning-AI/lightning-thunder/pull/2077
• add a use_hf option to benchmarking by @t-vi in https://github.com/Lightning-AI/lightning-thunder/pull/2154
• Bump pytest-xdist from 3.6.1 to 3.7.0 by @dependabot in https://github.com/Lightning-AI/lightning-thunder/pull/2164
• Update hypothesis requirement from ~=6.131.9 to ~=6.133.0 by @dependabot in https://github.com/Lightning-AI/lightning-thunder/pull/2165
• Update snowballstemmer requirement from <3 to <4 by @dependabot in https://github.com/Lightning-AI/lightning-thunder/pull/2168
• nvFuser Executor: Ensure cross-entropy loss fwd is not recomputed when computing bwd by @protonu in https://github.com/Lightning-AI/lightning-thunder/pull/2180
• Add parity check of shape/dtype/device of runtime and trace by @crcrpar in https://github.com/Lightning-AI/lightning-thunder/pull/2069
• Add docstrings for recipes by @KaelanDt in https://github.com/Lightning-AI/lightning-thunder/pull/2185
• sdpa_ex: relax test tolerances by @kshitij12345 in https://github.com/Lightning-AI/lightning-thunder/pull/2178
• removing nv_enable_embedding by @jjsjann123 in https://github.com/Lightning-AI/lightning-thunder/pull/2057
• Improve error reporting in benchmark job and add cleanup logic by @Borda in https://github.com/Lightning-AI/lightning-thunder/pull/2176
• Add mode flag to TorchCompileExecutor by @t-vi in https://github.com/Lightning-AI/lightning-thunder/pull/2188
• Use to_dtype and to_torch_dtype not _torch_to_thunder_dtype_map and _thunder_to_torch_dtype_map by @crcrpar in https://github.com/Lightning-AI/lightning-thunder/pull/2181
• use hf recipe in quickstart by @t-vi in https://github.com/Lightning-AI/lightning-thunder/pull/2191
• Remove kwarg construction from FusionDefinitionWrapper.__call__ by @IvanYashchuk in https://github.com/Lightning-AI/lightning-thunder/pull/1871
• Add tests for HFTransformers recipe with static cache by @KaelanDt in https://github.com/Lightning-AI/lightning-thunder/pull/2179
• bump: PyTorch to be latest 2.7.1 by @Borda in https://github.com/Lightning-AI/lightning-thunder/pull/2193
• prims.where ignores shape/device of pred if it's a CPU scalar tensor by @crcrpar in https://github.com/Lightning-AI/lightning-thunder/pull/2135
• add decomposition for repeat interleave by @t-vi in https://github.com/Lightning-AI/lightning-thunder/pull/2194
• fix traceback in with / try: finally: for Python 3.10 by @t-vi in https://github.com/Lightning-AI/lightning-thunder/pull/2195
• Use joint trace in transform_for_execution by @beverlylytle in https://github.com/Lightning-AI/lightning-thunder/pull/2102
• Handle proxy objects in the cuDNN SDPA checker by @kiya00 in https://github.com/Lightning-AI/lightning-thunder/pull/2073
• default to hf recipe in thunder.compile for hf models by @KaelanDt in https://github.com/Lightning-AI/lightning-thunder/pull/2199
• add autocast lookaside to hf recipe for tracing on meta device by @t-vi in https://github.com/Lightning-AI/lightning-thunder/pull/2200
• make test_networks.py not rely on HF downloads by @KaelanDt in https://github.com/Lightning-AI/lightning-thunder/pull/2202
• Add plugins documentation by @KaelanDt in https://github.com/Lightning-AI/lightning-thunder/pull/2207
• implement partial, avoid tuple addition, test partialmethod by @t-vi in https://github.com/Lightning-AI/lightning-thunder/pull/2209
• Add bitwise_left_shift and bitwise_right_shift by @crcrpar in https://github.com/Lightning-AI/lightning-thunder/pull/2210
• [thunderfx] Avoid split at Tensor.__eq__ by registering it in thunder.torch by @crcrpar in https://github.com/Lightning-AI/lightning-thunder/pull/2211
• fixed installing NCCL for CUDA by @Borda in https://github.com/Lightning-AI/lightning-thunder/pull/2208
• Enable ruff-check in pre-commit by @crcrpar in https://github.com/Lightning-AI/lightning-thunder/pull/2192
• unxfail passing test by @t-vi in https://github.com/Lightning-AI/lightning-thunder/pull/2220
• Add #2192 to .git-blame-ignore-revs by @crcrpar in https://github.com/Lightning-AI/lightning-thunder/pull/2219
• fix cache validity issue, tighten assert by @t-vi in https://github.com/Lightning-AI/lightning-thunder/pull/2223
• Avoid negative number rhs values to bitwise shift tests by @crcrpar in https://github.com/Lightning-AI/lightning-thunder/pull/2227

*…