diff --git a/optimum/intel/neural_compressor/trainer.py b/optimum/intel/neural_compressor/trainer.py
index 5f1caf2ae5..1923076bb5 100644
--- a/optimum/intel/neural_compressor/trainer.py
+++ b/optimum/intel/neural_compressor/trainer.py
@@ -56,6 +56,7 @@
 from transformers.training_args import ParallelMode, TrainingArguments
 from transformers.utils import (
     WEIGHTS_NAME,
+    is_accelerate_available,
     is_apex_available,
     is_sagemaker_mp_enabled,
     is_torch_tpu_available,
@@ -69,12 +70,31 @@
 from .configuration import INCConfig
 
 
+if is_accelerate_available():
+    from accelerate import __version__ as accelerate_version
+    from accelerate import skip_first_batches
+
+    if version.parse(accelerate_version) > version.parse("0.20.3"):
+        pass
+    DATA_SAMPLERS = [RandomSampler]
+    if version.parse(accelerate_version) > version.parse("0.23.0"):
+        from accelerate.data_loader import SeedableRandomSampler
+
+        DATA_SAMPLERS += [SeedableRandomSampler]
+
+    if is_deepspeed_available():
+        pass
+
+
 if is_apex_available():
     from apex import amp
 
 if is_sagemaker_mp_enabled():
     import smdistributed.modelparallel.torch as smp
 
+if is_torch_tpu_available(check_device=False):
+    import torch_xla.core.xla_model as xm
+
 
 if TYPE_CHECKING:
     from optimum.exporters.onnx import OnnxConfig