King Dedede
Reputation: 1010
How can I resume a training job in Sagemaker script mode?
In non-script mode for Sagemaker training jobs with Tensorflow, I was able to specify a checkpoint path in S3 with checkpoint_path.
However, in script mode this parameter is disabled.
How can I start from most recent checkpoint for a Sagemaker Tensorflow training job in Script mode?
Upvotes: 1
Views: 379
Answers (1)
lauren
Reputation: 513
with script mode, the parameter you're looking for is model_dir
docs: https://sagemaker.readthedocs.io/en/stable/using_tf.html#adapting-your-local-tensorflow-script
Upvotes: 1
Related Questions
- SageMaker's TensorFlow Training Job - how to change script archive path?
- How to run sagemaker processing and training job?
- AWS Sagemaker Pipelines throws a "No finished training job found associated with this estimator" after introducing a register step
- AWS Sagemaker training job stuck in progress state
- AWS Sagemaker Multiple Training Jobs
- error to run training job with aws Sagemaker
- How to train and deploy model in script mode on Sagemaker without using jupyter notebook instance (serverless)?
- How to schedule task in AWS Sagemaker to run training job
- SageMaker Script Mode Serving
- Continuous Training in Sagemaker