multi-host 使用iv-filter来进行federated feature selection，但是报错Host metrics not found in provided model,请问应该这么解决，配置跟报错如下

[Defu-csh]

binning_0: inputs: data: train_data: task_output_artifact: output_artifact_key: output_data producer_task: psi_0 parties: - role: guest party_id: - '9999' - role: host party_id: - '10001' - '10000' model: {} component_ref: hetero_feature_binning dependent_tasks: - psi_0 parameters: n_bins: 10 local_only: true transform_method: bin_idx method: bucket adjustment_factor: 0.5 use_anonymous: false relative_error: 1.0E-6 skip_metrics: false selection_0: inputs: data: train_data: task_output_artifact: output_artifact_key: output_data producer_task: psi_0 parties: - role: guest party_id: - '9999' - role: host party_id: - '10001' - '10000' model: input_models: task_output_artifact: - output_artifact_key: output_model producer_task: binning_0 parties: - role: guest party_id: - '9999' - role: host party_id: - '10001' - '10000' component_ref: hetero_feature_selection dependent_tasks: - psi_0 - binning_0 parameters: manual_param: keep_col: - x0 - x1 iv_param: filter_type: - threshold select_federated: true threshold: - 1.0 metrics: - iv host_filter_type: - threshold host_take_high: - true host_threshold: - 1.0 take_high: - true method: - iv statistic_param: filter_type: top_k threshold: 5 metrics: - max - mean use_anonymous: false keep_one: true

[nemirorox]

如需使用iv filter，上游binning需关闭local only，同时在selection组件注明接入模型和对应filter，可参考此样例配置：https://github.com/FederatedAI/FATE/blob/master/examples/pipeline/hetero_feature_selection/test_feature_selection_multi_host.py

[Defu-csh]

binning_0组件把参数local only改成false之后，跑三方还是报错

[Defu-csh]

@nemirorox