imporve log

Former-commit-id: a6abf375975ffea3d51e1b944c9855b5f62ffac8

src/llamafactory/data/data_utils.py

@@ -56,12 +56,12 @@ def merge_dataset(
         return all_datasets[0]
     elif data_args.mix_strategy == "concat":
         if data_args.streaming:
-            logger.warning_once("The samples between different datasets will not be mixed in streaming mode.")
+            logger.warning_rank0_once("The samples between different datasets will not be mixed in streaming mode.")
 
         return concatenate_datasets(all_datasets)
     elif data_args.mix_strategy.startswith("interleave"):
         if not data_args.streaming:
-            logger.warning_once("We recommend using `mix_strategy=concat` in non-streaming mode.")
+            logger.warning_rank0_once("We recommend using `mix_strategy=concat` in non-streaming mode.")
 
         return interleave_datasets(
             datasets=all_datasets,

src/llamafactory/data/loader.py

@@ -18,11 +18,10 @@ from typing import TYPE_CHECKING, Dict, Literal, Optional, Sequence, Union
 
 import numpy as np
 from datasets import DatasetDict, load_dataset, load_from_disk
-from transformers.utils.versions import require_version
 
 from ..extras import logging
 from ..extras.constants import FILEEXT2TYPE
-from ..extras.misc import has_tokenized_data
+from ..extras.misc import check_version, has_tokenized_data
 from .aligner import align_dataset
 from .data_utils import merge_dataset, split_dataset
 from .parser import get_dataset_list
@@ -84,7 +83,7 @@ def _load_single_dataset(
         raise NotImplementedError(f"Unknown load type: {dataset_attr.load_from}.")
 
     if dataset_attr.load_from == "ms_hub":
-        require_version("modelscope>=1.11.0", "To fix: pip install modelscope>=1.11.0")
+        check_version("modelscope>=1.11.0", mandatory=True)
         from modelscope import MsDataset  # type: ignore
         from modelscope.utils.config_ds import MS_DATASETS_CACHE  # type: ignore
 
@@ -103,7 +102,7 @@ def _load_single_dataset(
             dataset = dataset.to_hf_dataset()
 
     elif dataset_attr.load_from == "om_hub":
-        require_version("openmind>=0.8.0", "To fix: pip install openmind>=0.8.0")
+        check_version("openmind>=0.8.0", mandatory=True)
         from openmind import OmDataset  # type: ignore
         from openmind.utils.hub import OM_DATASETS_CACHE  # type: ignore
 

src/llamafactory/data/mm_plugin.py

@@ -73,10 +73,14 @@ class BasePlugin:
         Validates if this model accepts the input modalities.
         """
         if len(images) != 0 and self.image_token is None:
-            raise ValueError("This model does not support image input.")
+            raise ValueError(
+                "This model does not support image input. Please check whether the correct `template` is used."
+            )
 
         if len(videos) != 0 and self.video_token is None:
-            raise ValueError("This model does not support video input.")
+            raise ValueError(
+                "This model does not support video input. Please check whether the correct `template` is used."
+            )
 
     def _preprocess_image(self, image: "ImageObject", **kwargs) -> "ImageObject":
         r"""

src/llamafactory/data/template.py

@@ -15,10 +15,10 @@
 from dataclasses import dataclass
 from typing import TYPE_CHECKING, Dict, List, Optional, Sequence, Tuple, Union
 
-from transformers.utils.versions import require_version
 from typing_extensions import override
 
 from ..extras import logging
+from ..extras.misc import check_version
 from .data_utils import Role
 from .formatter import EmptyFormatter, FunctionFormatter, StringFormatter, ToolFormatter
 from .mm_plugin import get_mm_plugin
@@ -365,7 +365,7 @@ def get_template_and_fix_tokenizer(tokenizer: "PreTrainedTokenizer", data_args:
             raise ValueError(f"Template {data_args.template} does not exist.")
 
     if template.mm_plugin.__class__.__name__ != "BasePlugin":
-        require_version("transformers>=4.45.0", "To fix: pip install transformers>=4.45.0")
+        check_version("transformers>=4.45.0")
 
     if data_args.train_on_prompt and template.efficient_eos:
         raise ValueError("Current template does not support `train_on_prompt`.")