checkpoint model on first step callback (#2906)

* checkpoint model on first step callback

* remove debug

* add test cases; update existing tests not to save on first step

* move test out of solo

* delete

* default to False

* typo

tests/e2e/patched/test_phi_multipack.py

@@ -54,6 +54,7 @@ class TestPhiMultipack(unittest.TestCase):
                 "eval_steps": 3,
                 "save_steps": 4,
                 "bf16": "auto",
+                "save_first_step": False,
             }
         )
 
@@ -105,6 +106,7 @@ class TestPhiMultipack(unittest.TestCase):
                 "eval_steps": 3,
                 "save_steps": 4,
                 "bf16": "auto",
+                "save_first_step": False,
             }
         )