From 673325dea7309c9bb022527c0f773a0804b9f2b9 Mon Sep 17 00:00:00 2001 From: Daniel Han Date: Mon, 17 Feb 2025 20:21:20 -0800 Subject: [PATCH] Update rl_replacements.py --- unsloth/models/rl_replacements.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/unsloth/models/rl_replacements.py b/unsloth/models/rl_replacements.py index d8a1c63715..ee57055a00 100644 --- a/unsloth/models/rl_replacements.py +++ b/unsloth/models/rl_replacements.py @@ -237,7 +237,7 @@ def grpo_trainer_compute_loss(function_name, function): ref_per_token_logps, per_token_logps, input_ids, completion_mask, self.beta, advantages, bsz, ) accumulated_loss, accumulated_completion_length, accumulated_mean_kl = grpo_accumulated_loss( - self, input_ids, logits_to_keep, completion_mask, advantages, n_chunks = 1, + self, _input_ids, logits_to_keep, completion_mask, advantages, n_chunks = 1, ) print("loss", loss, accumulated_loss) print("completion_length", completion_length, accumulated_completion_length)