fix import error of flash attn

modeling_chatglm.py

@@ -21,16 +21,20 @@ from transformers.modeling_outputs import (
     SequenceClassifierOutputWithPast,
 )
 from transformers.modeling_utils import PreTrainedModel
-from transformers.utils import logging, is_torch_npu_available, is_flash_attn_greater_or_equal_2_10, \
-    is_flash_attn_2_available
+from transformers.utils import logging, is_torch_npu_available
 from transformers.generation.logits_process import LogitsProcessor
 from transformers.generation.utils import LogitsProcessorList, StoppingCriteriaList, GenerationConfig, ModelOutput
 
 from .configuration_chatglm import ChatGLMConfig
 
-if is_flash_attn_2_available():
-    from flash_attn import flash_attn_func, flash_attn_varlen_func
-    from flash_attn.bert_padding import index_first_axis, pad_input, unpad_input  # noqa
+try:
+    from transformers.utils import is_flash_attn_greater_or_equal_2_10, is_flash_attn_2_available
+    if is_flash_attn_2_available():
+        from flash_attn import flash_attn_func, flash_attn_varlen_func
+        from flash_attn.bert_padding import index_first_axis, pad_input, unpad_input  # noqa
+except:
+    pass
+
 
 # flags required to enable jit fusion kernels