run workflow_config_lightgbm_multi_freq.yaml: KeyError: 'PRED'

Bug Description

when i run .\examples\benchmarks\LightGBM\workflow_config_lightgbm_multi_freq.yaml , i got error:

[9944:MainThread](2024-01-24 20:50:10,472) INFO - qlib.workflow - [recorder.py:374] - Fail to log the uncommitted code of $CWD(D:\Project\playground\quant\qlib\workflow_config_lightgbm_multi_freq) when run git diff --cached.
[9944:MainThread](2024-01-24 20:50:55,250) INFO - qlib.timer - [log.py:128] - Time cost: 41.668s | Loading data Done
[9944:MainThread](2024-01-24 20:50:55,253) INFO - qlib.timer - [log.py:128] - Time cost: 0.002s | DropnaLabel Done
D:\Anaconda\envs\py39\lib\site-packages\pyqlib-0.9.1.99-py3.9-win-amd64.egg\qlib\data\dataset\processor.py:324: SettingWithCopyWarning:
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
df[cols] = df[cols].groupby("datetime", group_keys=False).apply(self.zscore_func)
[9944:MainThread](2024-01-24 20:50:55,265) INFO - qlib.timer - [log.py:128] - Time cost: 0.012s | CSZScoreNorm Done
[9944:MainThread](2024-01-24 20:50:55,266) INFO - qlib.timer - [log.py:128] - Time cost: 0.016s | fit & process data Done
[9944:MainThread](2024-01-24 20:50:55,267) INFO - qlib.timer - [log.py:128] - Time cost: 41.685s | Init data Done
Training until validation scores don't improve for 50 rounds
[20] train's l2: 0.996655 valid's l2: 0.996655
[40] train's l2: 0.996655 valid's l2: 0.996655
Early stopping, best iteration is:
[1] train's l2: 0.996655 valid's l2: 0.996655
[9944:MainThread](2024-01-24 20:50:55,519) INFO - qlib.timer - [log.py:128] - Time cost: 0.186s | waiting async_log Done
[9944:MainThread](2024-01-24 20:50:55,522) ERROR - qlib.workflow - [utils.py:41] - An exception has been raised[KeyError: 'PRED'].
File "D:\Anaconda\envs\py39\Scripts\qrun-script.py", line 33, in
sys.exit(load_entry_point('pyqlib==0.9.1.99', 'console_scripts', 'qrun')())
File "D:\Anaconda\envs\py39\lib\site-packages\pyqlib-0.9.1.99-py3.9-win-amd64.egg\qlib\workflow\cli.py", line 72, in run
fire.Fire(workflow)
File "d:\anaconda\envs\py39\lib\site-packages\fire\core.py", line 141, in Fire
component_trace = _Fire(component, args, parsed_flag_args, context, name)
File "d:\anaconda\envs\py39\lib\site-packages\fire\core.py", line 475, in _Fire
component, remaining_args = _CallAndUpdateTrace(
File "d:\anaconda\envs\py39\lib\site-packages\fire\core.py", line 691, in _CallAndUpdateTrace
component = fn(*varargs, **kwargs)
File "D:\Anaconda\envs\py39\lib\site-packages\pyqlib-0.9.1.99-py3.9-win-amd64.egg\qlib\workflow\cli.py", line 66, in workflow
recorder = task_train(config.get("task"), experiment_name=experiment_name)
File "D:\Anaconda\envs\py39\lib\site-packages\pyqlib-0.9.1.99-py3.9-win-amd64.egg\qlib\model\trainer.py", line 127, in task_train
exe_task(task_config)
File "D:\Anaconda\envs\py39\lib\site-packages\pyqlib-0.9.1.99-py3.9-win-amd64.egg\qlib\model\trainer.py", line 56, in exe_task
task_config = fill_placeholder(task_config, placehorder_value)
File "D:\Anaconda\envs\py39\lib\site-packages\pyqlib-0.9.1.99-py3.9-win-amd64.egg\qlib\utils_init.py", line 929, in fill_placeholder
now_item[key] = get_item_from_obj(config, m.groupdict()["name_path"])
File "D:\Anaconda\envs\py39\lib\site-packages\pyqlib-0.9.1.99-py3.9-win-amd64.egg\qlib\utils_init.py", line 874, in get_item_from_obj
cur_cfg = cur_cfg[k]
KeyError: 'PRED'

To Reproduce

Steps to reproduce the behavior:
use command:
qrun workflow_config_lightgbm_multi_freq.yaml

I used daily data and 1min freq data downloaded with qlib GetData().

Environment

• Qlib version: 0.9.1.99
• Python version: py39
• OS : windows 10

Start with examples/benchmarks/LightGBM/workflow_config_lightgbm_multi_freq.yaml and reproduce it with qrun. Trace placeholder handling through qlib/model/trainer.py and qlib/utils/init.py, especially fill_placeholder and get_item_from_obj, to identify why PRED is missing. Done means the workflow runs without KeyError: 'PRED'.