From 5200e67cd2a30d62a978b27f4d390ff0ba4f8ce7 Mon Sep 17 00:00:00 2001
From: "pre-commit-ci[bot]"
 <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Date: Thu, 1 Jan 2026 05:13:11 +0000
Subject: [PATCH] [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci
---
 unsloth/models/rl_replacements.py | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

diff --git a/unsloth/models/rl_replacements.py b/unsloth/models/rl_replacements.py
index c8231c315d..5e079335ae 100644
--- a/unsloth/models/rl_replacements.py
+++ b/unsloth/models/rl_replacements.py
@@ -861,13 +861,19 @@ def grpo_trainer_compute_loss(function_name, function):
                 else torch.tensor(0.0, device = self.model.device)
             )
             self._metrics[mode]["sampling/importance_sampling_ratio/min"].append(
-                self.accelerator.gather(min_importance_sampling_ratio).nan_to_num(nan=float('inf')).min().item()
+                self.accelerator.gather(min_importance_sampling_ratio)
+                .nan_to_num(nan = float("inf"))
+                .min()
+                .item()
             )
             self._metrics[mode]["sampling/importance_sampling_ratio/mean"].append(
                 self.accelerator.gather(mean_importance_sampling_ratio).nanmean().item()
             )
             self._metrics[mode]["sampling/importance_sampling_ratio/max"].append(
-                self.accelerator.gather(max_importance_sampling_ratio).nan_to_num(nan=float('-inf')).max().item()
+                self.accelerator.gather(max_importance_sampling_ratio)
+                .nan_to_num(nan = float("-inf"))
+                .max()
+                .item()
             )
 
         return loss