Quickstart¶
Run your first GIANT inference on a whole-slide image.
Prerequisites¶
- Installation completed
- OpenAI or Anthropic API key configured in
.env - A WSI file (
.svs,.tiff, or.ndpiformat)
Download a Test Slide¶
For testing, download a small WSI from OpenSlide's test data:
mkdir -p data/test
curl -L -o data/test/CMU-1-Small-Region.svs \
https://openslide.cs.cmu.edu/download/openslide-testdata/Aperio/CMU-1-Small-Region.svs
Run Inference¶
Basic Usage¶
# Activate environment and load API keys
source .venv/bin/activate
source .env
# Run GIANT on a WSI with a question
giant run data/test/CMU-1-Small-Region.svs \
-q "What type of tissue is shown in this slide?"
Expected Output¶
CLI Options¶
# Use Anthropic instead of OpenAI
giant run slide.svs -q "Question?" --provider anthropic
# Limit navigation steps
giant run slide.svs -q "Question?" --max-steps 5
# Set a cost budget (USD)
giant run slide.svs -q "Question?" --budget-usd 0.10
# Save the navigation trajectory
giant run slide.svs -q "Question?" --output trajectory.json
# Multiple runs with majority voting
giant run slide.svs -q "Question?" --runs 3
# JSON output for scripting
giant run slide.svs -q "Question?" --json
Understanding the Output¶
GIANT returns:
| Field | Description |
|---|---|
answer |
The model's response to your question |
cost |
Total API cost in USD |
turns |
Number of navigation steps taken |
agreement |
(with --runs > 1) Fraction of runs that agreed |
Visualize Navigation¶
After running with --output, visualize the agent's trajectory:
This opens an interactive HTML viewer showing:
- Initial thumbnail with axis guides
- Each cropped region the agent examined
- The agent's reasoning at each step
- Final answer
Next Steps¶
- First Benchmark - Run on real benchmark data
- Algorithm Explanation - Understand how GIANT navigates
- CLI Reference - All command options