do_predict后，生成的result里面除了label，predict字段,是否可以增加原始问题的字段？

[wickedjava]

Reminder

• I have read the README and searched the existing issues.

System Info

使用的llama-factory master branch

Reproduction

model

model_name_or_path: /home/work/Qwen1.5-7B-Chat
adapter_name_or_path: /home/work/Qwen1.5-7B-Chat/lora

method

stage: sft
do_predict: true
finetuning_type: lora

dataset

dataset: identity,alpaca_en_demo
dataset_dir: /home/work/git/LLaMA-Factory-main/data
template: qwen
cutoff_len: 1024
max_samples: 50
overwrite_cache: true
preprocessing_num_workers: 16

output

output_dir: /home/work/Qwen1.5-7B-Chat/lora/predict/v1
overwrite_output_dir: true

eval

per_device_eval_batch_size: 1
predict_with_generate: true
ddp_timeout: 180000000

Expected behavior

result 里面只有label和predict字段，能否增加原始的问题字段？

{"label": "Hello! I am {{name}}, an AI assistant developed by {{author}}. How can I assist you today?", "predict": "Hello! How can I help you today?"}

Others

No response

[baiyin]

看到已经新增了 ”prompt“ 字段，但还能不能带自定义的字段，比如 ”id" 字段到 result 中？