Jason Senthil

add num_sanity_eval_steps arg to fit

1

Differential Revision: D53205637

make updates to migration doc

1

Differential Revision: D52591234

fix oss docs

11

Differential Revision: D52377030

fix _retrieve_checkpoint_dirpaths

2

Summary: # Context For directories containing `_` in other parts of the name besides `epoch_0_step_0` (ex `tmp/fjad_213/epoch_0_step_0`), `_retrieve_checkpoint_dirpaths` can raise errors as it splits on `_` assuming underscore only appears...

add async flag to CSVWriter

1

Summary: Adds `async_write` flag to `CSVLogger` to make the `log()` call unblocking Reviewed By: galrotem Differential Revision: D51505159

add keep_latest_n checkpoint arg in TSS

1

Reviewed By: galrotem Differential Revision: D51048523

share `apply_strategy` method between autounit and autopredictunit

2

Summary: ## Context: Both `AutoUnit` and `AutoPredictUnit` use the same code block to apply the strategy on the module and check for any incompatibilties: ``` if strategy: if isinstance(strategy, str):...

replace warning w/ hard error in AutoUnit's fsdp + torch compile

4

Reviewed By: daniellepintz, rohan-varma Differential Revision: D47473960

update checkpoint docs to use unit instead of auto unit

2

Reviewed By: ananthsub Differential Revision: D48245705

fix torch compile tests in OSS

8

Differential Revision: D47803991