diff --git a/README.md b/README.md
index 250069a1..9e76b58e 100644
--- a/README.md
+++ b/README.md
@@ -483,13 +483,13 @@ huggingface-cli login
 ```bash
 git clone --depth 1 https://github.com/hiyouga/LLaMA-Factory.git
 cd LLaMA-Factory
-pip install -e ".[torch,metrics]"
+pip install -e ".[torch,metrics]" --no-build-isolation
 ```
 
 Extra dependencies available: torch, torch-npu, metrics, deepspeed, liger-kernel, bitsandbytes, hqq, eetq, gptq, aqlm, vllm, sglang, galore, apollo, badam, adam-mini, qwen, minicpm_v, modelscope, openmind, swanlab, quality
 
 > [!TIP]
-> Use `pip install --no-deps -e .` to resolve package conflicts.
+> Use `pip install -e . --no-deps --no-build-isolation` to resolve package conflicts.
 
 <details><summary>Setting up a virtual environment with <b>uv</b></summary>
 
diff --git a/README_zh.md b/README_zh.md
index baa2b20c..90ad5a0d 100644
--- a/README_zh.md
+++ b/README_zh.md
@@ -470,13 +470,13 @@ huggingface-cli login
 ```bash
 git clone --depth 1 https://github.com/hiyouga/LLaMA-Factory.git
 cd LLaMA-Factory
-pip install -e ".[torch,metrics]"
+pip install -e ".[torch,metrics]" --no-build-isolation
 ```
 
 可选的额外依赖项：torch、torch-npu、metrics、deepspeed、liger-kernel、bitsandbytes、hqq、eetq、gptq、aqlm、vllm、sglang、galore、apollo、badam、adam-mini、qwen、minicpm_v、modelscope、openmind、swanlab、quality
 
 > [!TIP]
-> 遇到包冲突时，可使用 `pip install --no-deps -e .` 解决。
+> 遇到包冲突时，可使用 `pip install -e . --no-deps --no-build-isolation` 解决。
 
 <details><summary>使用 <b>uv</b> 构建虚拟环境</summary>
 
diff --git a/assets/wechat.jpg b/assets/wechat.jpg
index 0cc68158..62849d99 100644
Binary files a/assets/wechat.jpg and b/assets/wechat.jpg differ
diff --git a/assets/wechat_npu.jpg b/assets/wechat_npu.jpg
index e3e58fea..33367385 100644
Binary files a/assets/wechat_npu.jpg and b/assets/wechat_npu.jpg differ
diff --git a/scripts/vllm_infer.py b/scripts/vllm_infer.py
index ad785253..b91634bb 100644
--- a/scripts/vllm_infer.py
+++ b/scripts/vllm_infer.py
@@ -17,8 +17,8 @@ import json
 from typing import Optional
 
 import fire
-from transformers import Seq2SeqTrainingArguments
 from tqdm import tqdm
+from transformers import Seq2SeqTrainingArguments
 
 from llamafactory.data import get_dataset, get_template_and_fix_tokenizer
 from llamafactory.extras.constants import IGNORE_INDEX
diff --git a/src/llamafactory/data/mm_plugin.py b/src/llamafactory/data/mm_plugin.py
index cd987672..a237b907 100644
--- a/src/llamafactory/data/mm_plugin.py
+++ b/src/llamafactory/data/mm_plugin.py
@@ -63,8 +63,8 @@ if is_transformers_version_greater_than("4.49.0"):
     except ImportError:
         try:
             # If that fails, try importing from the new location
-            from transformers.video_utils import make_batched_videos
             from transformers.image_utils import make_flat_list_of_images
+            from transformers.video_utils import make_batched_videos
         except ImportError:
             raise ImportError(
                 "Could not import make_batched_videos and make_flat_list_of_images. "
diff --git a/src/llamafactory/train/pt/workflow.py b/src/llamafactory/train/pt/workflow.py
index 85158c2c..17ea604b 100644
--- a/src/llamafactory/train/pt/workflow.py
+++ b/src/llamafactory/train/pt/workflow.py
@@ -84,14 +84,15 @@ def run_pt(
                     perplexity = math.exp(metrics[f"eval_{key}_loss"])
                 except OverflowError:
                     perplexity = float("inf")
+
                 metrics[f"eval_{key}_perplexity"] = perplexity
         else:
             try:
                 perplexity = math.exp(metrics["eval_loss"])
             except OverflowError:
                 perplexity = float("inf")
-            metrics["eval_perplexity"] = perplexity
 
+            metrics["eval_perplexity"] = perplexity
 
         trainer.log_metrics("eval", metrics)
         trainer.save_metrics("eval", metrics)