Refactor/chat ray

applications/Chat/examples/ray/1mmt_dummy.py → applications/Chat/benchmarks/ray/1mmt_dummy.py

@@ -63,7 +63,8 @@ def model_fn():
         critic_cfg = AutoConfig.from_pretrained(args.critic_pretrain)
         actor = get_actor_from_args(args.model, config=actor_cfg).requires_grad_(False).half().cuda()
         critic = get_critic_from_args(args.critic_model, config=critic_cfg).requires_grad_(False).half().cuda()
-        reward_model = get_reward_model_from_args(args.critic_model, config=critic_cfg).requires_grad_(False).half().cuda()
+        reward_model = get_reward_model_from_args(args.critic_model,
+                                                  config=critic_cfg).requires_grad_(False).half().cuda()
         if args.initial_model_quant_ckpt is not None and args.model == 'llama':
             # quantize initial model
             with low_resource_init(), no_init_weights():

applications/Chat/examples/ray/mmmt_dummy.py → applications/Chat/benchmarks/ray/mmmt_dummy.py

@@ -63,7 +63,8 @@ def model_fn():
         critic_cfg = AutoConfig.from_pretrained(args.critic_pretrain)
         actor = get_actor_from_args(args.model, config=actor_cfg).requires_grad_(False).half().cuda()
         critic = get_critic_from_args(args.critic_model, config=critic_cfg).requires_grad_(False).half().cuda()
-        reward_model = get_reward_model_from_args(args.critic_model, config=critic_cfg).requires_grad_(False).half().cuda()
+        reward_model = get_reward_model_from_args(args.critic_model,
+                                                  config=critic_cfg).requires_grad_(False).half().cuda()
         if args.initial_model_quant_ckpt is not None and args.model == 'llama':
             # quantize initial model
             with low_resource_init(), no_init_weights():

applications/Chat/coati/models/lora.py

@@ -62,7 +62,7 @@ def T(w):
             # Make sure that the weights are not merged
             if self.r > 0:
                 if not hasattr(self, "lora_A") or not hasattr(self, "lora_B"):
-                    # csric: temporary fix
+                    # FIXME(csric): temporary fix
                     self.lora_A = nn.Parameter(self.weight.new_empty((self.r, self.in_features)))
                     self.lora_B = nn.Parameter(self.weight.new_empty((self.out_features, self.r)))
                     self.reset_parameters()