Summary
We are excited to release this new version of the neuraloperator library! It brings many improvements, including new architectures (SFNO, GNO, GINO), many improvements to existing ones, out-of-the-box super resolution, super-evaluation and incremental training.
All models can now be easily saved and loaded and we provide a lightweight trainer compatible with all our neuraloperators. Head to the examples for some sample code, and to the API for a full documentation!
What's Changed
- Refactor MLP config by @JeanKossaifi in #143
- Adds super-resolution to FNO by @JeanKossaifi in #147
- Adds ADA_IN norm by @JeanKossaifi in #148
- Adds SFNO by @JeanKossaifi in #150
- Fix missing parameters output_scaling_factor by @sleepyeye in #159
- Finodev by @ashiq24 in #152
- quick_avoid by @ashiq24 in #173
- Add low-precision to TFNO by @crwhite14 in #172
- fix SFNO example by @crwhite14 in #177
- fix comma in readme file by @gegewen in #179
- Liftproj mod to mlp by @btolooshams in #182
- MLP additional statement to check for scenario where n_layers=1 by @btolooshams in #183
- marge the update on fno mlp help description by @btolooshams in #191
- Update guide on Fourier neural operator by @devzhk in #156
- Spectrum analysis of datasets. by @Robertboy18 in #193
- adding flag option to only pad the last dim by @btolooshams in #185
- Minor error in L#602 in fno.py by @ImanLiao in #194
- gino by @kovachki in #195
- Fix mlp nonlinearity by @ziqi-ma in #197
- Reformat
layers/
directory withblack
by @m4e7 in #199 - Use lifting channels in FNO.lifting if it is passed by @dhpitt in #196
- removed dead lines in FNOGNO by @dhpitt in #203
- Reformat
datasets/
directory withblack
by @m4e7 in #205 - docstring for FNOGNO by @dhpitt in #202
- Fix Python 3.6 f-string compatibility and condense documentation for FNO classes by @m4e7 in #209
- Split
preactivation
fromFNOBlock.forward()
by @m4e7 in #214 - Padding correction by @ashiq24 in #218
- general trainer class for GINO and NO by @dhpitt in #215
- Padding correction by @ashiq24 in #220
- fix example Trainer API calls by @dhpitt in #219
- Refactor rescaling in skips as transform in the Spectral Conv by @JeanKossaifi in #217
- Further simplification + UNO fix by @JeanKossaifi in #221
- Sht correction by @ashiq24 in #222
- Fix loss signatures to build doc by @dhpitt in #224
- fix small bug in the WandB logger callback by @dhpitt in #232
- Adding 4D prediction only, no nested fno by @gegewen in #225
- Revert 4D_FNO changes until they are properly tested by @dhpitt in #235
- Remove torch_scatter and torch_cluster from CI pipeline's dependencies by @dhpitt in #233
- Model checkpointing by @dhpitt in #234
- Updates to documentation and callback docstrings by @dhpitt in #237
- index dropout moduleList by @dhpitt in #239
- fix syntax error and add index.rst by @dhpitt in #240
- Fix typo in checkpoint init by @rybchuk in #241
- Bug fixes and unit testing for Callbacks by @dhpitt in #242
- Refactors SpectralConv for simpler FNO by @JeanKossaifi in #244
- BaseModel: adds checkpointing, versioning, safeguards by @JeanKossaifi in #257
- Add Burger's dataset and PINO by @crwhite14 in #256
- Enable transform wrappers by @JeanKossaifi in #254
- Update tensor_dataset.py by @slanthaler in #260
- Update to the checkpoint callback and test by @dhpitt in #258
- fix domain_padding to accept list (e.g., [0,0,1]) in addition to sc… by @btolooshams in #263
- Move to DataProcessor API by @dhpitt in #262
- Fix navier stokes preprocessor bug by @dhpitt in #265
- Fixes to make
DataProcessor
code doc build by @dhpitt in #266 - Add AutoML via Optuna by @crwhite14 in #243
- fixing the horizontal_skips_map construction, it was not going through by @btolooshams in #267
- Updates to saving and loading models by @dhpitt in #268
New Contributors
- @sleepyeye made their first contribution in #159
- @crwhite14 made their first contribution in #172
- @gegewen made their first contribution in #179
- @btolooshams made their first contribution in #182
- @devzhk made their first contribution in #156
- @Robertboy18 made their first contribution in #193
- @ImanLiao made their first contribution in #194
- @kovachki made their first contribution in #195
- @ziqi-ma made their first contribution in #197
- @m4e7 made their first contribution in #199
- @dhpitt made their first contribution in #196
- @rybchuk made their first contribution in #241
- @slanthaler made their first contribution in #260
Full Changelog: 0.2.0...0.3.0