fixes to accelerator so that iterable pretraining datasets work (#1759)

* fixes to accelerator so that iterable pretraining datasets work

* fix the pretraining test params

* split batches, not dispatch batches needs to be set

* update c4 datasets

* set epochs in pretrain config test

* need to set both split_batches and dispatch_batches to false for pretraining

* fix bool val in comment

tests/test_packed_pretraining.py

@@ -24,7 +24,7 @@ class TestPretrainingPacking(unittest.TestCase):
     def test_packing_stream_dataset(self):
         # pylint: disable=duplicate-code
         dataset = load_dataset(
-            "c4",
+            "allenai/c4",
             "en",
             streaming=True,
         )["train"]
@@ -33,7 +33,7 @@ class TestPretrainingPacking(unittest.TestCase):
             {
                 "pretraining_dataset": [
                     {
-                        "path": "c4",
+                        "path": "allenai/c4",
                         "name": "en",
                         "type": "pretrain",
                     }