r - 无法将模型或工件记录到 Mlflow
问题描述
我有一个 MLFlow 远程服务器,我可以从 rstudio 记录一些东西,比如:
mlflow_log_param("param1", 5)
mlflow_log_param("param2", 5)
mlflow_log_metric("foo", 1)
mlflow_log_metric("foo", 2)
mlflow_log_metric("foo", 3)
但是当我尝试记录以下内容时:
writeLines("Hello world!", "output.txt")
mlflow_log_artifact("output.txt")
我有这个错误:
2020/04/01 21:45:27 INFO mlflow.store.artifact.cli: Logged artifact from local file output.txt to artifact_path=None
Root URI: ./mlruns/12/3256cfd3cd1b44b99334040bd5c7c9ee/artifacts
当我尝试记录模型时:
mlflow_log_model(predictor, "model1")
我有下一个错误:
2020/04/01 21:56:44 INFO mlflow.store.artifact.cli: Logged artifact from local dir D:/user/AppData/Local/Temp/RtmpcBwDOP/model1 to artifact_path=model1
Root URI: ./mlruns/12/07d84e0f252a4248bb2473229297d318/artifacts
# A tibble: 0 x 0
Warning message:
In value[[3L]](cond) :
Logging model metadata to the tracking server has failed, possibly due to older server version. The model artifacts have been logged successfully. In addition to exporting model artifacts, MLflow clients 1.7.0 and above attempt to record model metadata to the tracking store. If logging to a mlflow server via REST, consider upgrading the server version to MLflow 1.7.0 or above.
服务器已更新,客户端也已更新。
解决方案
推荐阅读
- c# - 使用 Rosyln 将多个文件动态编译成程序集
- salesforce - 如何以编程方式对 LWC LightningDatatable 进行编辑?
- java - PROTOTYPE 的 ProxyMode 和 REQUEST 范围的区别
- angular - Angular 10:还在推荐 TSLint 或 ESLint 程序?
- ffmpeg - 有关 ffmpeg 命令的更多详细信息
- python - Python AttributeError:来自父函数的类继承
- c - 循环遍历字符串以查找不需要的字符
- javascript - 在 React 中放置 websocket 的位置
- react-native - KeyboardAwareScrollView 有额外的底部填充?
- python - 使用相似的键但不同的值更新字典/列表并在 python 中显示为报告