PiperOrigin-RevId: 512771693

PiperOrigin-RevId: 492530276

PiperOrigin-RevId: 489592044

PiperOrigin-RevId: 458550388

PiperOrigin-RevId: 452380386

PiperOrigin-RevId: 446234160

PiperOrigin-RevId: 422612138

PiperOrigin-RevId: 389019529

PiperOrigin-RevId: 385674527

PiperOrigin-RevId: 384478988

In nlp/train.py and vision/beta/train.py, certain flags are marked as required. Additionally, in certain functions, error messages are improved if a necessary flag is not specified, which is a fallback in case a file calling define_flags() does not mark the necessary flags are required. Previously if any of these flags were not specified, it would crash with a cryptic error message, making it hard to tell what went wrong.

In a subsequent change, I will mark flags as required in more files which call define_flags().

PiperOrigin-RevId: 381066985

PiperOrigin-RevId: 380858205

PiperOrigin-RevId: 380470496

PiperOrigin-RevId: 373623867

PiperOrigin-RevId: 368778443

PiperOrigin-RevId: 366879385

[core] Update BestCheckpointExporter to support dictionary with arbitrary depth (possibly coming from multitask eval).
Also added best checkpointexporter support to multitask.

PiperOrigin-RevId: 357584596

PiperOrigin-RevId: 352728348

We use kwargs for all arguments and enforce them should be kwargs.
Reorder args.

PiperOrigin-RevId: 346908658

PiperOrigin-RevId: 345054031

PiperOrigin-RevId: 342910130

PiperOrigin-RevId: 342179055

PiperOrigin-RevId: 342098346

When loss is NaN, the weights should also be NaN.

PiperOrigin-RevId: 341886095

PiperOrigin-RevId: 336795303

PiperOrigin-RevId: 336042908

Remove type annotation for params.
Make the trainer consume both model and task.

PiperOrigin-RevId: 334964198

PiperOrigin-RevId: 331861386

PiperOrigin-RevId: 327459481

PiperOrigin-RevId: 327459481

PiperOrigin-RevId: 327324127

PiperOrigin-RevId: 327324127

PiperOrigin-RevId: 326947485

PiperOrigin-RevId: 326947485

PiperOrigin-RevId: 326586473

PiperOrigin-RevId: 326586473