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

examples/llama-4/do-no-use-fa2/maverick-qlora-fsdp1.yaml

@@ -86,3 +86,5 @@ fsdp_config:
 special_tokens:
   pad_token: <|finetune_right_pad_id|>
   eos_token: <|eot|>
+
+# save_first_step: true  # uncomment this to validate checkpoint saving works with your config

examples/llama-4/do-no-use-fa2/scout-qlora-fsdp1.yaml

@@ -90,3 +90,5 @@ fsdp_config:
 special_tokens:
   pad_token: <|finetune_right_pad_id|>
   eos_token: <|eot|>
+
+# save_first_step: true  # uncomment this to validate checkpoint saving works with your config

examples/llama-4/do-no-use-fa2/scout-qlora-single-h100.yaml

@@ -83,3 +83,5 @@ weight_decay: 0.0
 special_tokens:
   pad_token: <|finetune_right_pad_id|>
   eos_token: <|eot|>
+
+# save_first_step: true  # uncomment this to validate checkpoint saving works with your config

examples/llama-4/do-no-use-fa2/scout-vision-qlora-fsdp.yaml

@@ -86,3 +86,5 @@ fsdp_config:
 special_tokens:
   pad_token: <|finetune_right_pad_id|>
   eos_token: <|eot|>
+
+# save_first_step: true  # uncomment this to validate checkpoint saving works with your config

examples/llama-4/scout-qlora-flexattn-fsdp2.yaml

@@ -84,3 +84,5 @@ fsdp_config:
 special_tokens:
   pad_token: <|finetune_right_pad_id|>
   eos_token: <|eot|>
+
+# save_first_step: true  # uncomment this to validate checkpoint saving works with your config

examples/llama-4/scout-qlora-single-h100-flex.yaml

@@ -82,3 +82,5 @@ weight_decay: 0.0
 special_tokens:
   pad_token: <|finetune_right_pad_id|>
   eos_token: <|eot|>
+
+# save_first_step: true  # uncomment this to validate checkpoint saving works with your config

examples/llama-4/scout-vision-qlora-fsdp2-flex.yaml

@@ -87,3 +87,5 @@ fsdp_config:
 special_tokens:
   pad_token: <|finetune_right_pad_id|>
   eos_token: <|eot|>
+
+# save_first_step: true  # uncomment this to validate checkpoint saving works with your config