初始化项目，由ModelHub XC社区提供模型

Model: openbmb/BitCPM-CANN-0.5B-unquantized
Source: Original Platform

example/run_sft.sh

@@ -0,0 +1,40 @@
+#!/bin/bash
+
+MODEL_PATH="/model/BitCPM-CANN-1B-unquantized"
+DATA_PATH="/dataset/HuggingFaceH4_ultrachat_200k/data/train_sft-00000-of-00003-a3ecf92756993583.parquet"
+OUTPUT_DIR="./output_sft"
+DS_CONFIG="./ds_config.json"
+
+NUM_GPUS=8
+BATCH_SIZE_PER_GPU=2
+GRAD_ACCUM_STEPS=1
+MAX_SEQ_LENGTH=8192
+
+export ASCEND_RT_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
+export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
+export DS_SKIP_CUDA_CHECK=1
+
+torchrun --nproc_per_node=$NUM_GPUS train_sft.py \
+    --model_name_or_path $MODEL_PATH \
+    --data_path $DATA_PATH \
+    --max_seq_length $MAX_SEQ_LENGTH \
+    --output_dir $OUTPUT_DIR \
+    --per_device_train_batch_size $BATCH_SIZE_PER_GPU \
+    --gradient_accumulation_steps $GRAD_ACCUM_STEPS \
+    --max_steps 100 \
+    --learning_rate 2e-5 \
+    --lr_scheduler_type cosine \
+    --warmup_ratio 0.2 \
+    --weight_decay 0.0 \
+    --logging_steps 2 \
+    --save_steps 500 \
+    --save_total_limit 3 \
+    --bf16 \
+    --deepspeed $DS_CONFIG \
+    --gradient_checkpointing \
+    --seed 42 \
+    --dataloader_num_workers 4 \
+    --report_to tensorboard \
+    --logging_dir /data/tensorboard/sft \
+    --train_on_prompt false \
+    --gradient_checkpointing_kwargs '{"use_reentrant": false}'