Train SegFormer for image segmentation.
Trains and evaluates SegFormer segmentation models and prepares them for inference through export and quantization.
When to use it
Use for training, evaluating, exporting, quantizing, or running inference with a TAO SegFormer model.
Give it a segmentation dataset and the requested model operation; it produces model results in a results directory.
What you provide
Requires Docker for the container-based workflow.
Requires NVIDIA Container Toolkit alongside Docker.
Requires at least one GPU with 16GB or more VRAM per GPU; recommends two GPUs and names V100 or A100.
Must run first for a standalone session not initialized by the TAO skill bank plugin, providing host preflight, credentials, and cross-skill discovery.
No additional actions listed.