fix bug: grpo train error for deepseek model

[aacedar]

PR type

use deepseek-ai/deepseek-coder-6.7b-base model run grpo algorithm， will find template_meta_prefix error： issue4808
we will find the error in swift/llm/template/base.py _swift_encode()，the concret code as following：

if self.template_meta.is_post_system or not system:

    prefix = template_meta.prefix

else:
    prefix = template_meta.system_prefix
self._concat_context_list(prefix, res_context_list, res_context_types, system=system)

PR information

debug the code will find the prefix value is [[32013]], the token of the prefix befor encode is ‘<｜begin▁of▁sentence｜>’.
the prompt join the prfix value and user input(str)，will raise error, as the issue4808

solution:
we just modify the file swift/trainers/rlhf_trainer/grpo_trainer.py, as this pr:

test script as following, just replace your_own_path in --model, --output_dir, you can test：

    CUDA_VISIBLE_DEVICES=7 \
    swift rollout \
      --model your_own_path/deepseek-ai/deepseek-coder-6.7b-base \ 
     > logs/log-grpo-rollout.log 2>&1 &

    CUDA_VISIBLE_DEVICES=0,1,2,3,4,5 \
    NPROC_PER_NODE=6 \
    swift rlhf \
      --rlhf_type grpo \
      --model your_own_path/deepseek-ai/deepseek-coder-6.7b-base \
      --reward_funcs accuracy \
      --use_vllm true \
      --vllm_mode server \
      --vllm_server_host 127.0.0.1 \
      --vllm_server_port 8000 \
      --train_type full \
      --torch_dtype bfloat16 \
      --dataset AI-MO/NuminaMath-TIR#1000 \
      --split_dataset_ratio 0 \
      --max_completion_length 512 \
      --num_train_epochs 1 \
      --per_device_train_batch_size 2 \
      --learning_rate 1e-6 \
      --gradient_accumulation_steps 2 \
      --save_total_limit 2 \
      --logging_steps 1 \
      --deepspeed zero2 \
      --max_length 4096 \
      --warmup_ratio 0.05 \
      --dataloader_num_workers 2 \
      --dataset_num_proc 4 \
      --num_generations 6 \
      --temperature 0.9 \
      --top_p 0.9 \
      --top_k 50 \
      --log_completions true \
      --num_iterations 1 \
      --beta 0.01 \
      --output_dir ./ds-base-grpo \
      --report_to tensorboard \
      > logs/log-grpo-test.log 2>&1 &

[hjh0119]

Please do not use Chinese comments.

[aacedar]

Please do not use Chinese comments.

has updated

[hjh0119]

plz resolve the conflicts

[aacedar]

plz resolve the conflicts

done