v0.8.2

[justinxzhao]

What's Changed

• int: Rename original combiner_registry to combiner_config_registry, update decorator name by @ksbrar in int: Rename original combiner_registry to combiner_config_registry, update decorator name #3516
• Add mechanic to override default values for generation during model.predict() by @justinxzhao in Add mechanic to override default values for generation during model.predict() #3520
• [feat] Support for numeric date feature inputs by @jeffkinnison in [feat] Support for numeric date feature inputs #3517
• Add new sythesized response column for text output features during postprocessing by @arnavgarg1 in Add new sythesized response column for text output features during postprocessing #3521
• Disable flaky twitter bots dataset loading test. by @justinxzhao in Disable flaky twitter bots dataset loading test. #3439
• Add test that verifies that the generation config passed in at model.predict() is used correctly. by @justinxzhao in Add test that verifies that the generation config passed in at model.predict() is used correctly. #3523
• Move loss metric to same device as inputs by @Infernaught in Move loss metric to same device as inputs #3522
• Add comment about batch size tuning by @arnavgarg1 in Add comment about batch size tuning #3526
• Ensure user sets backend to local w/ quantization by @Infernaught in Ensure user sets backend to local w/ quantization #3524
• README: Update LLM fine-tuning config by @arnavgarg1 in README: Update LLM fine-tuning config #3530
• Revert "Ensure user sets backend to local w/ quantization (Ensure user sets backend to local w/ quantization #3524)" by @tgaddair in Revert "Ensure user sets backend to local w/ quantization (#3524)" #3531
• Revert "Ensure user sets backend to local w/ quantization" for release-0.8 branch and upgrade version to 0.8.1.post1 by @justinxzhao in Revert "Ensure user sets backend to local w/ quantization" for release-0.8 branch and upgrade version to 0.8.1.post1 #3532
• Improve observability during LLM inference by @arnavgarg1 in Improve observability during LLM inference  #3536
• [bug] Pin pydantic to < 2.0 by @jeffkinnison in [bug] Pin pydantic to < 2.0 #3537
• [bug] Support preprocessing datetime.date date features by @jeffkinnison in [bug] Support preprocessing datetime.date date features #3534
• Remove obsolete prompt tuning example. by @justinxzhao in Remove obsolete prompt tuning example. #3540
• Add Ludwig 0.8 notebook to the README by @arnavgarg1 in Add Ludwig 0.8 notebook to the README #3542
• Add effective_batch_size to auto-adjust gradient accumulation by @tgaddair in Add effective_batch_size to auto-adjust gradient accumulation #3533
• Refactor evaluation metrics to support decoded generated text metrics like BLEU and ROUGE. by @justinxzhao in Refactor evaluation metrics to support decoded generated text metrics like BLEU and ROUGE. #3539
• Fix sequence generator test. by @justinxzhao in Fix sequence generator test. #3546
• DRAFT: Revert "Add Cosine Annealing LR scheduler as a decay method (Add Cosine Annealing LR scheduler as a decay method #3507)" by @justinxzhao in Revert "Add Cosine Annealing LR scheduler as a decay method (#3507)" #3545
• Set default max_sequence_length to None for LLM text input/output features by @arnavgarg1 in Set default max_sequence_length to None for LLM text input/output features #3547
• Add skip_all_evaluation as a mechanic to skip all evaluation. by @justinxzhao in Add skip_all_evaluation as a mechanic to skip all evaluation. #3543
• Roll-forward with fixes: Fix interaction between scheduler.step() and gradient accumulation steps, refactor schedulers to use LambdaLR, and add cosine annealing LR scheduler as a decay method. by @justinxzhao in Roll-forward with fixes: Fix interaction between scheduler.step() and gradient accumulation steps, refactor schedulers to use LambdaLR, and add cosine annealing LR scheduler as a decay method. #3555
• fix: Move model to the correct device for eval by @jeffkinnison in fix: Move model to the correct device for eval #3554
• Report loss in tqdm to avoid log spam by @tgaddair in Report loss in tqdm to avoid log spam #3559
• Wrap each metric update in try/except. by @justinxzhao in Wrap each metric update in try/except. #3562
• Move DDP model to device if it hasn't been wrapped yet by @tgaddair in Move DDP model to device if it hasn't been wrapped yet #3566
• ensure that there are enough colors to match the score index in visua… by @thelinuxkid in ensure that there are enough colors to match the score index in visua… #3560
• Pin Transformers to 4.31.0 by @arnavgarg1 in Pin Transformers to 4.31.0 #3569

New Contributors

• @thelinuxkid made their first contribution in ensure that there are enough colors to match the score index in visua… #3560

Full Changelog: v0.8.1...v0.8.2

---

This discussion was created from the release v0.8.2.