sagemaker.train

SageMaker Python SDK Train Module.

Modules

agent_rft_job	AgentRFTJob — wrapper around sagemaker-core Job for AgentRFT job category.
aws_batch
base_trainer	Base trainer providing shared fine-tuning workflow logic for SageMaker trainers.
common	Common types and options shared across SageMaker training modules.
configs	DEPRECATED: This module has been moved to sagemaker.core.training.configs
constants	DEPRECATED: This module has been moved to sagemaker.core.training.constants
container_drivers	Sagemaker modules container drivers directory.
cpt_trainer	CPTTrainer — Continued Pre-Training on foundation models using SageMaker HyperPod.
custom_agent_lambda	CustomAgentLambda — Lambda-based agent environment for Agentic RFT.
data_mixing_config	Configuration for blending customer training data with Nova curated datasets.
defaults	This module contains logic for setting defaults in ModelTrainer.
distributed	Distributed module.
dpo_trainer	DPO (Direct Preference Optimization) trainer for SageMaker fine-tuning.
evaluate	SageMaker Model Evaluation Module.
local
model_trainer	ModelTrainer class module.
multi_turn_rl_trainer	MultiTurnRLTrainer — trainer for Agentic Reinforcement Fine-Tuning (Multi-Turn RL) jobs.
recipe_resolver	Recipe resolution with 3-level override precedence for Nova model training.
remote_function	DEPRECATED: This module has been moved to sagemaker.core.remote_function
rft	SageMaker RFT SDK - Integration library for multi-turn RL training platform.
rlaif_trainer	RLAIF (Reinforcement Learning from AI Feedback) trainer for SageMaker fine-tuning.
rlvr_trainer	RLVR (Reinforcement Learning from Verifiable Rewards) trainer for SageMaker fine-tuning.
sft_trainer	SFT (Supervised Fine-Tuning) trainer for SageMaker fine-tuning.
sm_recipes
templates	Templates module.
tuner	Placeholder docstring
types	Types module.
utils	Utils module.