aws/sagemaker-training-toolkit

Issues

• Issue when training in local mode with huggingface training container

  #193 opened a year ago by ojturner
  1

• Unable to install sagemaker-training on Windows

  #110 opened 3 years ago by martinlyra
  3

• SageMaker Local Mode does not Inject default environment variables

  #107 opened 3 years ago by edgBR
  9

• Silent Failure if custom image puts something into /opt/ml/code

  #222 opened 3 months ago by njbrake
  0

• SageMaker training toolkit reorders hyperparameters

  #221 opened 3 months ago by vsimkus
  0

• Add 'ml.p5.48xlarge' as a supported instance for SM_EFA_NCCL_INSTANCES.

  #219 opened 3 months ago by andjsmi
  0

• Extend documentation regarding distributed training for own Docker containers.

  #218 opened 3 months ago by marseller
  0

• No support for Python 3.10

  #129 opened 2 years ago by peter-wimsey
  9

• Failed to parse string hyperparameter

  #66 opened 4 years ago by uwaisiqbal
  6

• Validate smddprun() fails with file not found error on AL2023

  #213 opened 8 months ago by jimmyrigby94
  1

• Get region with ENV var

  #207 opened 10 months ago by austinmw
  0

• Invalid dash-separated options for description-file

  #206 opened a year ago by wickeat
  0

• Training Job "Successful" despite failing due to 100% disk usage

  #204 opened a year ago by david-waterworth
  0

• ModuleNotFoundError: Sagemaker only copies entry_point file to /opt/ml/code/ instead of the holy-cloned source code

  #200 opened a year ago by celsofranssa
  0

• P5 instance support

  #191 opened a year ago by haozhx23
  0

• Support CodeArtifact repositories for installing Python packages

  #167 opened a year ago by humanzz
  0

• Adding sys.path to PYTHONPATH breaks virtual environments

  #186 opened a year ago by pdveenstra
  0

• Deepspeed Launcher

  #184 opened 2 years ago by anupam-dewan
  0

• Bash Command ENTRYPOINT Expects `train` argument

  #65 opened 4 years ago by uwaisiqbal
  8

• Publish wheels to PyPI

  #174 opened 2 years ago by hajapy
  0

• Passing SIGTERM to entrypoint to be able to handle SPOT failures gracefully in user-code

  #173 opened 2 years ago by croth1
  0

• Remove magic strings for attributes like instance type

  #164 opened 2 years ago by vishwakaria
  0

• Python 3.6 unsupported [bug/question]

  #161 opened 2 years ago by adamwrobel-ext-gd
  1

• Mpi mode sets all nodes to the same SM_CURRENT_HOST

  #158 opened 2 years ago by verdimrc
  0

• Custom MPI options doesn't override the flags

  #70 opened 4 years ago by ChaiBapchya
  3

• Version 4.1.4 fails to install because of the missing protobuf dependency

  #132 opened 2 years ago by szafranek
  2

• Hyperparameters not shell escaped

  #128 opened 3 years ago by bstriner
  0

• Pass SIGTERM to training script to stop training

  #125 opened 3 years ago by bstriner
  0

• Pass in environment variables for Estimator training job

  #97 opened 4 years ago by PiercingDan
  3

• Hyperparameters and other cmd arguments are not passed to shell entrypoint in tensorflow > 2.4

  #115 opened 3 years ago by unoebauer
  2

• Arguments not always accessible when using bash script for training job

  #119 opened 3 years ago by marcelgwerder
  3

• Model and output files do not get saved to S3 when training own model

  #82 opened 4 years ago by fiocam
  5

• Feature: Allow user script override of failure reason file

  #111 opened 3 years ago by athewsey
  1

• Should to_cmd_args pass complex types through json.dumps instead of str?

  #117 opened 3 years ago by croth1
  1

• Add Custom_Overrides flag.

  #113 opened 3 years ago by mathephysicist
  1

• Custom_Overrides

  #112 opened 3 years ago by mathephysicist
  0

• How to save non-model artifacts from a container (output_data_dir)

  #80 opened 4 years ago by diegodebrito
  2

• How does sagemaker-training-toolkit complement sagemaker-python-sdk?

  #105 opened 4 years ago by yanhong-zhao-ef
  0

• Don't set `sagemaker_s3_output` via hyperparameter

  #104 opened 4 years ago by samuel-massinon
  0

• Add gcc package requirment

  #94 opened 4 years ago by timorkal
  0

• how to pass predictor script for running batch transformers.

  #93 opened 4 years ago by AnkurShukla85
  0

• SageMaker Endpoint stuck at “Creating”

  #92 opened 4 years ago by vas610
  0

• Sagemaker Fails to download code from S3

  #85 opened 4 years ago by uwaisiqbal
  0

• Which aws service is the most suitable/used to launch a scheduled training job?

  #84 opened 4 years ago by david-fortini
  0

• Enable functional test for mpi

  #83 opened 4 years ago by ChaiBapchya
  0

• Enhance UX for training

  #77 opened 4 years ago by ehsanmok
  0

• Bash script mode support across all estimators

  #75 opened 4 years ago by ehsanmok
  1

• Unable to run a Tensorflow Estimator.

  #72 opened 4 years ago by nectario
  3

• Pass args to training script entrypoint for MPI-based Distributed training

  #69 opened 4 years ago by ChaiBapchya
  3

• Using the "typing" package causes conflicts in Python 3.7

  #60 opened 5 years ago by danielRicardo
  2