Revert "checkpoint model on first step callback (#2906)"

This reverts commit 10ba1622f7.

examples/mistral/bigstral-ds-zero3.yaml

@@ -53,5 +53,3 @@ special_tokens:
   eos_token: "<|im_end|>"
 tokens:
   - "<|im_start|>"
-
-# save_first_step: true  # uncomment this to validate checkpoint saving works with your config

examples/mistral/config.yml

@@ -43,5 +43,3 @@ evals_per_epoch: 4
 saves_per_epoch: 1
 weight_decay: 0.0
 special_tokens:
-
-# save_first_step: true  # uncomment this to validate checkpoint saving works with your config

examples/mistral/lora-mps.yml

@@ -64,5 +64,3 @@ evals_per_epoch: 4
 saves_per_epoch: 1
 weight_decay: 0.0
 special_tokens:
-
-# save_first_step: true  # uncomment this to validate checkpoint saving works with your config

examples/mistral/lora.yml

@@ -64,5 +64,3 @@ evals_per_epoch: 4
 saves_per_epoch: 1
 weight_decay: 0.0
 special_tokens:
-
-# save_first_step: true  # uncomment this to validate checkpoint saving works with your config

examples/mistral/mistral-dpo-qlora.yml

@@ -80,5 +80,3 @@ weight_decay: 0.0
 special_tokens:
   bos_token: "<|im_start|>"
   eos_token: "<|im_end|>"
-
-# save_first_step: true  # uncomment this to validate checkpoint saving works with your config

examples/mistral/mistral-qlora-fsdp.yml

@@ -74,5 +74,3 @@ fsdp_config:
   fsdp_state_dict_type: FULL_STATE_DICT
   fsdp_auto_wrap_policy: TRANSFORMER_BASED_WRAP
 special_tokens:
-
-# save_first_step: true  # uncomment this to validate checkpoint saving works with your config

examples/mistral/mistral-qlora-orpo.yml

@@ -69,5 +69,3 @@ evals_per_epoch: 4
 saves_per_epoch: 1
 weight_decay: 0.0
 special_tokens:
-
-# save_first_step: true  # uncomment this to validate checkpoint saving works with your config

examples/mistral/mistral-small-3.1-24B-lora.yml

@@ -56,5 +56,3 @@ evals_per_epoch: 1
 saves_per_epoch: 1
 weight_decay: 0.0
 special_tokens:
-
-# save_first_step: true  # uncomment this to validate checkpoint saving works with your config

examples/mistral/mixtral-8x22b-qlora-fsdp.yml

@@ -72,5 +72,3 @@ fsdp_config:
   fsdp_state_dict_type: FULL_STATE_DICT
   fsdp_auto_wrap_policy: TRANSFORMER_BASED_WRAP
 special_tokens:
-
-# save_first_step: true  # uncomment this to validate checkpoint saving works with your config

examples/mistral/mixtral-qlora-fsdp.yml

@@ -77,5 +77,3 @@ fsdp_config:
   fsdp_forward_prefetch: false
   fsdp_backward_prefetch: BACKWARD_PRE
 special_tokens:
-
-# save_first_step: true  # uncomment this to validate checkpoint saving works with your config

examples/mistral/mixtral.yml

@@ -81,5 +81,3 @@ saves_per_epoch: 1
 deepspeed: deepspeed_configs/zero2.json
 weight_decay: 0.0
 special_tokens:
-
-# save_first_step: true  # uncomment this to validate checkpoint saving works with your config

examples/mistral/mixtral_22.yml

@@ -51,5 +51,3 @@ special_tokens:
   eos_token: "<|im_end|>"
 tokens:
   - "<|im_start|>"
-
-# save_first_step: true  # uncomment this to validate checkpoint saving works with your config

examples/mistral/qlora.yml

@@ -64,5 +64,3 @@ evals_per_epoch: 4
 saves_per_epoch: 1
 weight_decay: 0.0
 special_tokens:
-
-# save_first_step: true  # uncomment this to validate checkpoint saving works with your config