Troubleshooting（故障排除）¶
This page contains tips for troubleshooting errors that you may encounter while using Model Studio. If you have an error that you cannot resolve with this guide, report an issue to Palantir Support.
No models trained¶
The No models trained error indicates that all model types were excluded when configuring your model studio. This can happen by excluding the models directly, or by using the hyperparameter field. To resolve this, set the hyperparameter field to {}, or reduce the number of excluded models. Refer to the time series forecasting, classification, or regression trainer documentation, depending on the trainer being used by your model studio.
Out of memory¶
Due to how datasets are stored in Foundry, you may run into out of memory (OOM) errors if you did not properly scale your memory to fit the dataset. Datasets produced in Foundry tend to be highly compressed, meaning that datasets may take up more memory when uncompressed. Provisioning more memory may also unlock general performance gains, as parallelized workers within the trainer can operate more efficiently. Learn more about configuring compute resources.
Frequency of train data is not provided and cannot be inferred¶
The Frequency of train_data is not provided and cannot be inferred error occurs when the time series trainer is unable to determine the frequency of your data when setting the resample configuration frequency argument to auto. To resolve this, disable resampling or manually set the resampling frequency. Refer to the time series forecasting trainer parameters for more information.
中文翻译¶
故障排除¶
本页面包含使用 Model Studio 时可能遇到的错误故障排除技巧。如果遇到本指南无法解决的问题，请向 Palantir 支持团队提交问题报告。
未训练任何模型¶
No models trained（未训练任何模型）错误表示在配置 Model Studio 时所有模型类型均被排除。这可能是直接排除了模型，或通过超参数字段（hyperparameter field）导致的。要解决此问题，请将超参数字段设置为 {}，或减少排除的模型数量。请根据 Model Studio 使用的训练器类型，参考时间序列预测、分类或回归训练器的文档。
内存不足¶
由于数据集在 Foundry 中的存储方式，如果未根据数据集规模合理分配内存，可能会遇到内存不足（OOM）错误。Foundry 生成的数据集通常经过高度压缩，这意味着解压后数据集可能占用更多内存。增加内存分配还可能提升整体性能，因为训练器中的并行工作进程可以更高效地运行。了解更多关于配置计算资源的信息。
未提供训练数据频率且无法推断¶
当时间序列训练器在将重采样配置频率参数（resample configuration frequency argument）设置为 auto 时无法确定数据频率，会出现 Frequency of train_data is not provided and cannot be inferred（未提供训练数据频率且无法推断）错误。要解决此问题，请禁用重采样或手动设置重采样频率。更多信息请参考时间序列预测训练器的参数文档。