| name | aggregate |
| description | Aggregate experiment results from work_dirs into a combined CSV and optionally plot a bubble chart. Use when the user wants to collect, summarize, filter, or visualize benchmark results for a dataset. |
When to use / what to ask
Ask the user:
- Which dataset to aggregate (e.g.
ETTh1, electricity)
- Optional filters (e.g.
pred_len=96, model~Linear, mse<=0.5 — operators =, !=, <, >, <=, >=, ~ substring)
- Whether to also plot a bubble chart
Commands
uv run python tool/aggregate_results.py --dataset <dataset> [--filter "pred_len=96,model~Linear"]
uv run python tool/tsf.py aggregate-plot --dataset <name> --pred-len <len>
For a fair leaderboard (TFB-style), add --collapse --aggregate mean --null-threshold 0.3 — collapses per-seed runs into one row per (model, pred_len) and excludes models missing on too many cells (dropped models are logged).
Notes
- Scans
work_dirs/<dataset>/*/performance.csv and profile.csv; profile columns (latency, params, VRAM) require evaluation.enable_profile = true in the run config.
- Column selection (
--perf-fields, --prof-fields, --metric-cols) supports all recorded metrics: mse, mae, corr, rse, wape, smape, mase.
Reference
Full flags: docs/en/aggregate-results.md; chart options: docs/en/plot-bubble.md (or the plot skill).