Pretrain transforms (#1261)

* wip for pretraining/iterable data with arbitrary prompt strategies

* more fixes, wip

* more fixes for custom pretraining

* iterable ds wrapper not needed

* remove extra features

* chore: lint

* update pretraning example yml

* fix order for partials

* fixup for tests

examples/tiny-llama/pretrain.yml

@@ -12,6 +12,7 @@ max_steps: 200
 pretraining_dataset:
   path: c4
   name: en
+  type: pretrain
 dataset_prepared_path:
 val_set_size: 0.0
 output_dir: ./model-out