ValueError: could not convert string to float

[Checunmily]

Hello, recently I tried xgb.cv and got an error when assigning save_best=True in xgb.callback.EarlyStopping().
I just used the demo privided at https://xgboost.readthedocs.io/en/latest/python/examples/cross_validation.html
Here is the full code I used:

import os

import numpy as np

import xgboost as xgb

# load data in do training
CURRENT_DIR = os.path.dirname(__file__)
dtrain = xgb.DMatrix(
    os.path.join(CURRENT_DIR, "../data/agaricus.txt.train?format=libsvm")
)
param = {"max_depth": 2, "eta": 1, "objective": "binary:logistic"}
num_round = 2

print("running cross validation, disable standard deviation display and save best")
# do cross validation, this will print result out as
# [iteration]  metric_name:mean_value
res = xgb.cv(
    param,
    dtrain,
    num_boost_round=10,
    nfold=5,
    metrics={"error"},
    seed=0,
    callbacks=[
        xgb.callback.EvaluationMonitor(show_stdv=False),
        xgb.callback.EarlyStopping(3, save_best=True),
    ],
)
print(res)

Finally I will got a

ValueError: could not convert string to float: '(0.001228171312096744, 0.0010411026299985337)'

When I tracing back, I found the function

def best_score(self) -> float:
    """Get best_score."""
    return float(cast(float, self.cvfolds[0].bst.attr("best_score")))

at

~/.conda/envs/sklearn-env/lib/python3.12/site-packages/xgboost/training.py", line 271

will return a float but

self.cvfolds[0].bst.attr("best_score")

just give a str like

(0.001228171312096744, 0.0010411026299985337)

When I trying to skip this step I got another error:

File "~/.conda/envs/sklearn-env/lib/python3.12/site-packages/xgboost/callback.py", line 464, in after_training
    model = model[: best_iteration + 1]
            ~~~~~^^^^^^^^^^^^^^^^^^^^^^
TypeError: '_PackedBooster' object is not subscriptable

Did I miss something? How can I fix it? Thanks in advance.

[trivialfis]

Thank you for raising the issue, will try to fix it. The CV function needs some update.