label data type error in evaluation when using "multi" eval_type

[kaiwang0112006]

我在用自定义算法跑联邦下的bert，参考教程：https://fate.readthedocs.io/en/develop-1.11.1/tutorial/pipeline/nn_tutorial/Bert-example/#dataset-imdb-sentimental

其中有一些细节我已经做了修改，IMDB的示例数据已经跑通，所以我进一步想跑一下文本多分类，使用了一个医学数据集：https://www.kaggle.com/datasets/jpmiller/layoutlm， 已经处理成和IMDB一样的格式：

id, text, label
1, some text,0
...

用notebook提交任务：

import torch as t
from torch import nn
from pipeline import fate_torch_hook
from pipeline.component import HomoNN
from pipeline.backend.pipeline import PipeLine
from pipeline.component import Reader, Evaluation, DataTransform
from pipeline.interface import Data, Model
import torch
fate_torch_hook(t)


import os
fate_project_path = os.path.abspath('/data/prjects/fate')
guest_0 = 10000
host_1 = 10000
pipeline = PipeLine().set_initiator(role='guest', party_id=guest_0).set_roles(guest=guest_0, host=host_1,
                                                                              arbiter=guest_0)
data_0 = {"name": "imdb", "namespace": "experiment"}
data_path = '/data/projects/fate/examples/data/train_for_fl.csv'
pipeline.bind_table(name=data_0['name'], namespace=data_0['namespace'], path=data_path)
pipeline.bind_table(name=data_0['name'], namespace=data_0['namespace'], path=data_path)

reader_0 = Reader(name="reader_0")
reader_0.get_party_instance(role='guest', party_id=guest_0).component_param(table=data_0)
reader_0.get_party_instance(role='host', party_id=host_1).component_param(table=data_0)

reader_1 = Reader(name="reader_1")
reader_1.get_party_instance(role='guest', party_id=guest_0).component_param(table=data_0)
reader_1.get_party_instance(role='host', party_id=host_1).component_param(table=data_0)

from pipeline.component.homo_nn import DatasetParam, TrainerParam  
model = t.nn.Sequential(
    t.nn.CustModel(module_name='bert_', class_name='BertClassifier')
)

nn_component = HomoNN(name='nn_0',
                      model=model, 
                      loss=t.nn.BCELoss(),
                      optimizer = t.optim.Adam(lr=0.001, weight_decay=0.001),
                      dataset=DatasetParam(dataset_name='nlp_tokenizer', tokenizer_name_or_path="bert-base-uncased"),  # 使用自定义的dataset
                      trainer=TrainerParam(trainer_name='fedavg_trainer', epochs=2, batch_size=16, data_loader_worker=8, cuda="cuda:0"),
                      torch_seed=100 
                      )

pipeline.add_component(reader_0)
pipeline.add_component(reader_1)
pipeline.add_component(nn_component, data=Data(train_data=reader_0.output.data, validate_data=reader_1.output.data))
pipeline.add_component(Evaluation(name='eval_0', eval_type='multi'), data=Data(data=nn_component.output.data))
pipeline.compile()
pipeline.fit()

所以重点是把evaluation改成了“multi”，任务在evaluation报错了：

[ERROR] [2024-07-16 16:31:08,499] [202407161623325775800] [38874:139965857994560] - [task_executor.run] [line:266]: 'numpy.float64' object cannot be interpreted as an integer
Traceback (most recent call last):
File "/data/projects/fate/fateflow/python/fate_flow/worker/task_executor.py", line 210, in run
cpn_output = run_object.run(cpn_input)
File "/data/projects/fate/fate/python/federatedml/model_base.py", line 239, in run
self._run(cpn_input=cpn_input)
File "/data/projects/fate/fate/python/federatedml/model_base.py", line 315, in _run
this_data_output = func(*real_param)
File "/data/projects/fate/fate/python/federatedml/evaluation/evaluation.py", line 400, in fit
return self.callback_metric_data(
File "/data/projects/fate/fate/python/federatedml/evaluation/evaluation.py", line 1001, in callback_metric_data
self.__save_pr_curve(precision_recall, data_type)
File "/data/projects/fate/fate/python/federatedml/evaluation/evaluation.py", line 697, in __save_pr_curve
pos_recall_score, recall_cuts = self.__multi_class_label_padding(
File "/data/projects/fate/fate/python/federatedml/evaluation/evaluation.py", line 478, in __multi_class_label_padding
for i in range(label_num):
TypeError: 'numpy.float64' object cannot be interpreted as an integer

顺着报错信息我看到/data/projects/fate/fate/python/federatedml/evaluation/evaluation.py这个文件，核心是label_num是float不是int，因此我修改了evaluation.py， 将label_num强制转成int，任务没有再报错，这里是bug吗？

[sagewe]

方便的话可以继续追踪下label_num为什么会是numpy.float64