DEDUPED Hot AI repos. Real use cases. What people build. Get the Monday 5

How to find the right task and class names before you run with Total Segmentator

Use totalseg_info to print the available tasks and their classes without downloading weights.

Short answer: Install it with pip install TotalSegmentator; then run totalseg_info --list-tasks for a table of tasks with modality, license and number of classes; run totalseg_info --classes -ta total to see the class index to name mapping for one task. All steps.

Who it is forA developer or coding agent wiring Total Segmentator into a pipeline.

The problemGuessing task and class names wastes whole GPU runs.

The outcomeAn exact list of tasks and class names, as a table or as JSON.

What you need first

  • Python 3.10 or newer and PyTorch 2.0.0 or newer
  • A Nifti file, or a folder or zip of all DICOM slices of one patient, as input
  • A GPU for reasonable runtime; CPU works with the --fast or --roi_subset options
  • A free non-commercial license number for the non-open tasks and for totalseg_get_body_stats

Steps

  1. Install it with pip install TotalSegmentator
  2. Run totalseg_info --list-tasks for a table of tasks with modality, license and number of classes
  3. Run totalseg_info --classes -ta total to see the class index to name mapping for one task
  4. Run totalseg_info --json to get the full capability registry as JSON for your script

Prompt for your agent

Using wasserth/TotalSegmentator, run `totalseg_info --json` and write me a small Python helper that picks the right task and `--roi_subset` names for <the structures I care about>.

Caveat. Some subtasks need a license number set with totalseg_set_license, which is free for non-commercial use.

Use this in your AI agent

Ask your agent

Using the Deduped MCP, get the setup steps for wasserth/TotalSegmentator and "Find the right task and class names before you run"

Add Deduped to Claude Code once

claude mcp add --transport http deduped https://dedupedai.com/mcp

Cursor, ChatGPT, others

Share: X LinkedIn Hacker News Email Copy link

Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer