diff --git a/examples/mindspore/gemma/pretrain_gemma_2b_ms.sh b/examples/mindspore/gemma/pretrain_gemma_2b_ms.sh
new file mode 100644
index 0000000000000000000000000000000000000000..07c117a5876b056cb88e43a8266d5e9ec2a1491b
--- /dev/null
+++ b/examples/mindspore/gemma/pretrain_gemma_2b_ms.sh
@@ -0,0 +1,104 @@
+#!/bin/bash
+export CUDA_DEVICE_MAX_CONNECTIONS=1
+
+NPUS_PER_NODE=8
+MASTER_ADDR=localhost
+MASTER_PORT=6000
+NNODES=1
+NODE_RANK=0
+WORLD_SIZE=$(($NPUS_PER_NODE*$NNODES))
+
+CKPT_SAVE_DIR="your model save ckpt path"
+DATA_PATH="your data path"
+TOKENIZER_MODEL="your tokenizer path"
+CKPT_LOAD_DIR="your model ckpt path"
+TP=1
+PP=2
+
+DISTRIBUTED_ARGS="
+    --master_addr $MASTER_ADDR \
+    --node_rank $NODE_RANK \
+    --worker_num $WORLD_SIZE \
+    --local_worker_num $NPUS_PER_NODE \
+    --master_port $MASTER_PORT \
+    --log_dir=msrun_log \
+    --join=False \
+    --cluster_time_out=300 \
+    --bind_core=True \
+"
+
+
+GPT_ARGS="
+    --tensor-model-parallel-size ${TP} \
+    --pipeline-model-parallel-size ${PP} \
+    --sequence-parallel \
+    --use-mcore-models \
+
+    --use-fused-rmsnorm \
+    --use-fused-rotary-pos-emb \
+    --num-layers 18 \
+    --hidden-size 2048 \
+    --ffn-hidden-size 16384 \
+    --num-attention-heads 8 \
+    --group-query-attention \
+    --num-query-groups 1 \
+    --tokenizer-type PretrainedFromHF \
+    --tokenizer-name-or-path ${TOKENIZER_MODEL} \
+    --seq-length 8192 \
+    --max-position-embeddings 8192 \
+    --micro-batch-size 1 \
+    --global-batch-size 256 \
+    --kv-channels 256 \
+    --make-vocab-size-divisible-by 1 \
+    --lr 1.25e-6 \
+    --train-iters 2000 \
+    --lr-decay-style cosine \
+    --disable-bias-linear \
+    --attention-dropout 0.0 \
+    --init-method-std 0.01 \
+    --hidden-dropout 0.0 \
+    --position-embedding-type rope \
+    --normalization RMSNorm \
+    --add-rmsnorm-offset \
+    --norm-epsilon 1e-06 \
+    --geglu \
+    --input-embeds-norm \
+    --use-flash-attn \
+    --no-masked-softmax-fusion \
+    --attention-softmax-in-fp32 \
+    --min-lr 1.25e-7 \
+    --weight-decay 1e-1 \
+    --lr-warmup-fraction 0.01 \
+    --clip-grad 1.0 \
+    --adam-beta1 0.9 \
+    --adam-beta2 0.95 \
+    --initial-loss-scale 4096 \
+    --use-distributed-optimizer \
+    --no-gradient-accumulation-fusion \
+    --no-load-optim \
+    --no-load-rng \
+    --vocab-size 256000 \
+    --bf16
+"
+
+DATA_ARGS="
+    --data-path $DATA_PATH \
+    --split 100,0,0
+"
+
+OUTPUT_ARGS="
+    --log-interval 1 \
+    --save-interval 2000 \
+    --eval-interval 1000 \
+    --eval-iters 0 \
+"
+
+msrun $DISTRIBUTED_ARGS pretrain_gpt.py \
+    $GPT_ARGS \
+    $DATA_ARGS \
+    $OUTPUT_ARGS \
+    --distributed-backend nccl \
+    --ai-framework mindspore \
+    --load ${CKPT_LOAD_DIR} \
+    --save ${CKPT_SAVE_DIR} \
+    | tee logs/train_gemma_2b_mcore.log
\ No newline at end of file
diff --git a/examples/mindspore/gemma/pretrain_gemma_7b_ms.sh b/examples/mindspore/gemma/pretrain_gemma_7b_ms.sh
new file mode 100644
index 0000000000000000000000000000000000000000..bcce363b25e7c543957bff6ee4de162ba82d5ebf
--- /dev/null
+++ b/examples/mindspore/gemma/pretrain_gemma_7b_ms.sh
@@ -0,0 +1,101 @@
+#!/bin/bash
+export CUDA_DEVICE_MAX_CONNECTIONS=1
+
+NPUS_PER_NODE=8
+MASTER_ADDR=localhost
+MASTER_PORT=6000
+NNODES=1
+NODE_RANK=0
+WORLD_SIZE=$(($NPUS_PER_NODE*$NNODES))
+
+CKPT_SAVE_DIR="your model save ckpt path"
+DATA_PATH="your data path"
+TOKENIZER_MODEL="your tokenizer path"
+CKPT_LOAD_DIR="your model ckpt path"
+TP=8
+PP=1
+
+DISTRIBUTED_ARGS="
+    --master_addr $MASTER_ADDR \
+    --node_rank $NODE_RANK \
+    --worker_num $WORLD_SIZE \
+    --local_worker_num $NPUS_PER_NODE \
+    --master_port $MASTER_PORT \
+    --log_dir=msrun_log \
+    --join=False \
+    --cluster_time_out=300 \
+    --bind_core=True \
+"
+
+
+GPT_ARGS="
+    --tensor-model-parallel-size ${TP} \
+    --pipeline-model-parallel-size ${PP} \
+    --sequence-parallel \
+    --use-mcore-models \
+
+    --use-fused-rmsnorm \
+    --use-fused-rotary-pos-emb \
+    --num-layers 28 \
+    --hidden-size 3072 \
+    --ffn-hidden-size 24576 \
+    --num-attention-heads 16 \
+    --tokenizer-type PretrainedFromHF \
+    --tokenizer-name-or-path ${TOKENIZER_MODEL} \
+    --seq-length 8192 \
+    --max-position-embeddings 8192 \
+    --micro-batch-size 2 \
+    --global-batch-size 64 \
+    --kv-channels 256 \
+    --make-vocab-size-divisible-by 1 \
+    --lr 1.25e-6 \
+    --train-iters 2000 \
+    --lr-decay-style cosine \
+    --disable-bias-linear \
+    --attention-dropout 0.0 \
+    --init-method-std 0.01 \
+    --hidden-dropout 0.0 \
+    --position-embedding-type rope \
+    --normalization RMSNorm \
+    --add-rmsnorm-offset \
+    --norm-epsilon 1e-06 \
+    --geglu \
+    --input-embeds-norm \
+    --use-flash-attn \
+    --no-masked-softmax-fusion \
+    --attention-softmax-in-fp32 \
+    --min-lr 1.25e-7 \
+    --weight-decay 1e-1 \
+    --lr-warmup-fraction 0.01 \
+    --clip-grad 1.0 \
+    --adam-beta1 0.9 \
+    --adam-beta2 0.95 \
+    --initial-loss-scale 4096 \
+    --no-gradient-accumulation-fusion \
+    --no-load-optim \
+    --no-load-rng \
+    --vocab-size 256000 \
+    --bf16
+"
+
+DATA_ARGS="
+    --data-path $DATA_PATH \
+    --split 100,0,0
+"
+
+OUTPUT_ARGS="
+    --log-interval 1 \
+    --save-interval 2000 \
+    --eval-interval 1000 \
+    --eval-iters 0 \
+"
+
+msrun $DISTRIBUTED_ARGS pretrain_gpt.py \
+    $GPT_ARGS \
+    $DATA_ARGS \
+    $OUTPUT_ARGS \
+    --distributed-backend nccl \
+    --ai-framework mindspore \
+    --load ${CKPT_LOAD_DIR} \
+    --save ${CKPT_SAVE_DIR} \
+    | tee logs/train_gemma_7b_mcore.log
\ No newline at end of file
diff --git a/examples/mindspore/gemma2/pretrain_gemma2_27b_ms.sh b/examples/mindspore/gemma2/pretrain_gemma2_27b_ms.sh
new file mode 100644
index 0000000000000000000000000000000000000000..79f6093e0b5f2b6a9cb2968efe2115d56a0f4772
--- /dev/null
+++ b/examples/mindspore/gemma2/pretrain_gemma2_27b_ms.sh
@@ -0,0 +1,112 @@
+#!/bin/bash
+export CUDA_DEVICE_MAX_CONNECTIONS=1
+
+NPUS_PER_NODE=8
+MASTER_ADDR=localhost
+MASTER_PORT=6000
+NNODES=2
+NODE_RANK=0
+WORLD_SIZE=$(($NPUS_PER_NODE*$NNODES))
+
+CKPT_SAVE_DIR="your model save ckpt path"
+DATA_PATH="your data path"
+TOKENIZER_MODEL="your tokenizer path"
+CKPT_LOAD_DIR="your model ckpt path"
+
+TP=8
+PP=2
+
+DISTRIBUTED_ARGS="
+    --master_addr $MASTER_ADDR \
+    --node_rank $NODE_RANK \
+    --worker_num $WORLD_SIZE \
+    --local_worker_num $NPUS_PER_NODE \
+    --master_port $MASTER_PORT \
+    --log_dir=msrun_log \
+    --join=False \
+    --cluster_time_out=300 \
+    --bind_core=True \
+"
+
+
+GPT_ARGS="
+    --tensor-model-parallel-size ${TP} \
+    --pipeline-model-parallel-size ${PP} \
+    --sequence-parallel \
+    --spec mindspeed_llm.tasks.models.spec.gemma2_spec layer_spec \
+    --use-mcore-models \
+
+    --use-fused-rmsnorm \
+    --use-fused-rotary-pos-emb \
+    --gelu-tanh \
+    --post-norm \
+    --query-pre-attn-scalar 144 \
+    --output-logit-softcapping 30.0 \
+    --interleave-sliding-window 4096 \
+    --num-layers 46 \
+    --num-layer-list 20,26 \
+    --hidden-size 4608 \
+    --ffn-hidden-size 36864 \
+    --num-attention-heads 32 \
+    --tokenizer-type PretrainedFromHF \
+    --tokenizer-name-or-path ${TOKENIZER_MODEL} \
+    --seq-length 8192 \
+    --max-position-embeddings 8192 \
+    --micro-batch-size 1 \
+    --global-batch-size 64 \
+    --kv-channels 128 \
+    --group-query-attention \
+    --num-query-groups 16 \
+    --make-vocab-size-divisible-by 1 \
+    --lr 1.25e-6 \
+    --train-iters 2000 \
+    --lr-decay-style cosine \
+    --disable-bias-linear \
+    --attention-dropout 0.0 \
+    --init-method-std 0.01 \
+    --hidden-dropout 0.0 \
+    --position-embedding-type rope \
+    --normalization RMSNorm \
+    --add-rmsnorm-offset \
+    --norm-epsilon 1e-06 \
+    --input-embeds-norm \
+    --use-flash-attn \
+    --use-distributed-optimizer \
+    --no-masked-softmax-fusion \
+    --attention-softmax-in-fp32 \
+    --min-lr 1.25e-7 \
+    --weight-decay 1e-1 \
+    --lr-warmup-fraction 0.01 \
+    --clip-grad 1.0 \
+    --adam-beta1 0.9 \
+    --adam-beta2 0.95 \
+    --tokenizer-padding-side left \
+    --initial-loss-scale 4096 \
+    --no-gradient-accumulation-fusion \
+    --no-load-optim \
+    --no-load-rng \
+    --vocab-size 256000 \
+    --bf16
+"
+
+DATA_ARGS="
+    --data-path $DATA_PATH \
+    --split 100,0,0
+"
+
+OUTPUT_ARGS="
+    --log-interval 1 \
+    --save-interval 2000 \
+    --eval-interval 1000 \
+    --eval-iters 0 \
+"
+
+msrun $DISTRIBUTED_ARGS pretrain_gpt.py \
+    $GPT_ARGS \
+    $DATA_ARGS \
+    $OUTPUT_ARGS \
+    --distributed-backend nccl \
+    --ai-framework mindspore \
+    --load ${CKPT_LOAD_DIR} \
+    --save ${CKPT_SAVE_DIR} \
+    | tee logs/train_gemma2_27b_mcore.log
\ No newline at end of file
diff --git a/examples/mindspore/gemma2/pretrain_gemma2_9b_ms.sh b/examples/mindspore/gemma2/pretrain_gemma2_9b_ms.sh
new file mode 100644
index 0000000000000000000000000000000000000000..2cc82da0dc4e1b1e43b6aef2865e86729aa852d4
--- /dev/null
+++ b/examples/mindspore/gemma2/pretrain_gemma2_9b_ms.sh
@@ -0,0 +1,111 @@
+#!/bin/bash
+export CUDA_DEVICE_MAX_CONNECTIONS=1
+
+NPUS_PER_NODE=8
+MASTER_ADDR=localhost
+MASTER_PORT=6000
+NNODES=1
+NODE_RANK=0
+WORLD_SIZE=$(($NPUS_PER_NODE*$NNODES))
+
+CKPT_SAVE_DIR="your model save ckpt path"
+DATA_PATH="your data path"
+TOKENIZER_MODEL="your tokenizer path"
+CKPT_LOAD_DIR="your model ckpt path"
+
+TP=8
+PP=1
+
+DISTRIBUTED_ARGS="
+    --master_addr $MASTER_ADDR \
+    --node_rank $NODE_RANK \
+    --worker_num $WORLD_SIZE \
+    --local_worker_num $NPUS_PER_NODE \
+    --master_port $MASTER_PORT \
+    --log_dir=msrun_log \
+    --join=False \
+    --cluster_time_out=300 \
+    --bind_core=True \
+"
+
+
+GPT_ARGS="
+    --tensor-model-parallel-size ${TP} \
+    --pipeline-model-parallel-size ${PP} \
+    --sequence-parallel \
+    --spec mindspeed_llm.tasks.models.spec.gemma2_spec layer_spec \
+    --use-mcore-models \
+
+    --use-fused-rmsnorm \
+    --use-fused-rotary-pos-emb \
+    --gelu-tanh \
+    --post-norm \
+    --query-pre-attn-scalar 256 \
+    --output-logit-softcapping 30.0 \
+    --interleave-sliding-window 4096 \
+    --num-layers 42 \
+    --hidden-size 3584 \
+    --ffn-hidden-size 14336 \
+    --num-attention-heads 16 \
+    --tokenizer-type PretrainedFromHF \
+    --tokenizer-name-or-path ${TOKENIZER_MODEL} \
+    --seq-length 8192 \
+    --max-position-embeddings 8192 \
+    --micro-batch-size 1 \
+    --global-batch-size 64 \
+    --kv-channels 256 \
+    --group-query-attention \
+    --num-query-groups 8 \
+    --make-vocab-size-divisible-by 1 \
+    --lr 1.25e-6 \
+    --train-iters 2000 \
+    --lr-decay-style cosine \
+    --disable-bias-linear \
+    --attention-dropout 0.0 \
+    --init-method-std 0.01 \
+    --hidden-dropout 0.0 \
+    --position-embedding-type rope \
+    --normalization RMSNorm \
+    --add-rmsnorm-offset \
+    --norm-epsilon 1e-06 \
+    --input-embeds-norm \
+    --use-flash-attn \
+    --use-distributed-optimizer \
+    --no-masked-softmax-fusion \
+    --attention-softmax-in-fp32 \
+    --min-lr 1.25e-7 \
+    --weight-decay 1e-1 \
+    --lr-warmup-fraction 0.01 \
+    --clip-grad 1.0 \
+    --adam-beta1 0.9 \
+    --adam-beta2 0.95 \
+    --tokenizer-padding-side left \
+    --initial-loss-scale 4096 \
+    --no-gradient-accumulation-fusion \
+    --no-load-optim \
+    --no-load-rng \
+    --vocab-size 256000 \
+    --bf16
+"
+
+DATA_ARGS="
+    --data-path $DATA_PATH \
+    --split 100,0,0
+"
+
+OUTPUT_ARGS="
+    --log-interval 1 \
+    --save-interval 2000 \
+    --eval-interval 1000 \
+    --eval-iters 0 \
+"
+
+msrun $DISTRIBUTED_ARGS pretrain_gpt.py \
+    $GPT_ARGS \
+    $DATA_ARGS \
+    $OUTPUT_ARGS \
+    --distributed-backend nccl \
+    --ai-framework mindspore \
+    --load ${CKPT_LOAD_DIR} \
+    --save ${CKPT_SAVE_DIR} \
+    | tee logs/train_gemma2_9b_mcore.log
\ No newline at end of file