Update accelerate_ppo_trainer.py

CarperAI · Nov 15, 2023 · fd3d95b · fd3d95b
1 parent d283ee2
commit fd3d95b
Showing 1 changed file with 2 additions and 2 deletions.
diff --git a/trlx/trainer/accelerate_ppo_trainer.py b/trlx/trainer/accelerate_ppo_trainer.py
@@ -352,9 +352,9 @@ def make_experience(self, num_rollouts: int = 1024, iter_count: int = 0):  # noq
             scores_mask = scores != -np.inf
 
             if self.config.train.reward_only_in_main_process:
-                str_samples, str_prompts, str_outputs = self.decode(prompt_tensors, samples, append_eos_token=True)
+                _, _, str_outputs = self.decode(prompt_tensors, samples, append_eos_token=True)
             else:
-                str_samples, str_prompts, str_outputs = all_str_samples, all_str_prompts, all_str_outputs
+                str_outputs = all_str_outputs
 
             # Pad the sample outputs
             outputs = self.tokenizer(str_outputs).input_ids