Εnable reward model offloading option (deepspeedai#930)

* enable reward model offloading option

* fixed code formatting

* more formatting fixes

* Pre-commit formatting fix

---------

Co-authored-by: Olatunji Ruwase <[email protected]>
Co-authored-by: Logan Adams <[email protected]>
Co-authored-by: Logan Adams <[email protected]>
Signed-off-by: zhangsmallshark <[email protected]>

Author: 3 people

applications/DeepSpeed-Chat/dschat/rlhf/rlhf_engine.py

@@ -268,31 +268,22 @@ def _init_reward(self, critic_model_name_or_path):
             # If critic is ZeRO-3 then we use it for everything, otherwise assume we have enough memory
             zero_stage = 0
 
-        ds_config = get_eval_ds_config(offload=self.args.offload,
+        ds_config = get_eval_ds_config(offload=self.args.offload_reward_model,
                                        dtype=self.args.dtype,
                                        stage=zero_stage)
-        ds_config[
-            'train_micro_batch_size_per_gpu'] = self.args.per_device_training_batch_size
-        ds_config[
-            'train_batch_size'] = self.args.per_device_training_batch_size * torch.distributed.get_world_size(
-            ) * self.args.gradient_accumulation_steps
-
-        ds_eval_config = get_eval_ds_config(offload=False,
-                                            dtype=self.args.dtype,
-                                            stage=zero_stage)
 
         # We need to set train batch size and micro batch size here to pass the sanity check of DeepSpeed engine.
-        ds_eval_config[
+        ds_config[
             'train_micro_batch_size_per_gpu'] = self.args.per_device_training_batch_size
-        ds_eval_config[
+        ds_config[
             'train_batch_size'] = self.args.per_device_training_batch_size * torch.distributed.get_world_size(
             ) * self.args.gradient_accumulation_steps
 
         # Model
         reward_model = create_critic_model(
             model_name_or_path=critic_model_name_or_path,
             tokenizer=self.tokenizer,
-            ds_config=ds_eval_config,
+            ds_config=ds_config,
             num_padding_at_beginning=self.args.num_padding_at_beginning,
             rlhf_training=True,
             dropout=self.args.critic_dropout,

applications/DeepSpeed-Chat/training/step3_rlhf_finetuning/main.py

@@ -246,6 +246,9 @@ def parse_args():
         '--offload_reference_model',
         action='store_true',
         help='Enable ZeRO Offload techniques for reference model')
+    parser.add_argument('--offload_reward_model',
+                        action='store_true',
+                        help='Enable ZeRO Offload techniques for reward model')
     parser.add_argument(
         '--actor_zero_stage',
         type=int,