FATE 模型训练出现“TypeError: a bytes-like object is required, not 'NoneType'”

运行一个住房预测的案例，在训练模型的时候出现not 'NoneType'”。具体的过程如下

总共有4个文件分别是housing_1_eval、housing_1_train、housing_2_eval、housing_2_train

然后使用upload_host.json文件把housing_1_eval和housing_1_train上传到host方容器。具体的文件内容如下

然后使用upload_guest.json文件把housing_2_eval和housing_2_train上传至guest方容器。具体的文件内容如下

然后在host中的容器提交任务，

dag: parties:

party_id:
- '9999' role: guest
party_id:
- '10000' role: host
party_id:
- '10000' role: arbiter party_tasks: guest_9999: parties:
  - party_id:
    - '9999' role: guest tasks: reader_0: parameters: name: housing_train_guest namespace: experiment host_10000: parties:
  - party_id:
    - '10000' role: host tasks: reader_0: parameters: name: housing_train_host namespace: experiment stage: train tasks: reader_0: component_ref: reader parties:
    - party_id:
      - '9999' role: guest
    - party_id:
      - '10000' role: host stage: default binning_0: component_ref: hetero_feature_binning dependent_tasks:
  - scale_0 inputs: data: train_data: task_output_artifact: output_artifact_key: train_output_data producer_task: scale_0 model: {} parameters: adjustment_factor: 0.5 bin_col: null bin_idx: null category_col: null category_idx: null local_only: false method: quantile n_bins: 10 relative_error: 1.0e-06 skip_metrics: false split_pt_dict: null transform_method: null use_anonymous: false parties:
  - party_id:
    - '9999' role: guest
  - party_id:
    - '10000' role: host evaluation_0: component_ref: evaluation dependent_tasks:
  - lr_0 inputs: data: input_data: task_output_artifact: - output_artifact_key: train_output_data producer_task: lr_0 parties: - party_id: - '9999' role: guest parameters: default_eval_setting: binary label_column_name: null metrics: null predict_column_name: null parties:
  - party_id:
    - '9999' role: guest stage: default lr_0: component_ref: coordinated_lr dependent_tasks:
  - selection_0 inputs: data: train_data: task_output_artifact: output_artifact_key: train_output_data producer_task: selection_0 parties: - party_id: - - '9999' role: guest - party_id: - '10000' role: host model: {} parameters: batch_size: null early_stop: diff epochs: 10 floating_point_precision: 23 output_cv_data: true threshold: 0.5 tol: 0.0001 psi_0: component_ref: psi inputs: data: input_data: task_output_artifact: output_artifact_key: output_data producer_task: reader_0 parameters: {} parties:
  - party_id:
    - '9999' role: guest
  - party_id:
    - '10000' role: host stage: default scale_0: component_ref: feature_scale dependent_tasks:
  - psi_0 inputs: data: train_data: task_output_artifact: output_artifact_key: output_data producer_task: psi_0 model: {} parameters: feature_range: null method: standard scale_col: null scale_idx: null strict_range: true use_anonymous: false parties:
  - party_id:
    - '9999' role: guest
  - party_id:
'10000' role: host selection_0: component_ref: hetero_feature_selection dependent_tasks: - binning_0 - scale_0 inputs: data: train_data: task_output_artifact: output_artifact_key: train_output_data producer_task: scale_0 model: input_models: task_output_artifact: - output_artifact_key: output_model producer_task: binning_0 parameters: iv_param: filter_type: threshold metrics: iv threshold: 0.1 keep_one: true manual_param: null method: - iv select_col: null statistic_param: null use_anonymous: false parties: - party_id: - '9999' role: guest - party_id: - '10000' role: host schema_version: 2.0.0

然后在host容器中提交任务

[root@976cf52cd022 fate_flow]# flow job submit -c examples/lr/train_lr.yaml { "code": 0, "data": { "model_id": "202403122047397316310", "model_version": "0" }, "job_id": "202403122047397316310", "message": "success" }

在host（192.168.0.132）的Borad中没有错误输出，在guest（192.168.0.133）的Borad中出现“TypeError: a bytes-like object is required, not 'NoneType'”报错