Use TraceML With Your Stack

TraceML can be added to a normal PyTorch training loop or used through a framework-specific entry point. Pick the path that matches how your training job already runs.

Stack	TraceML entry point	Start here
Plain PyTorch loop	traceml.trace_step(...)	Quickstart
Hugging Face Trainer	TraceMLTrainerCallback	Hugging Face
PyTorch Lightning	traceml_ai.integrations.lightning.init() + TraceMLCallback	PyTorch Lightning
Ray Train	TraceMLTorchTrainer	Ray Train
W&B / MLflow	traceml.summary()	W&B / MLflow