Upload benchmark/tabarena/SUBMISSION_DRAFT.md with huggingface_hub

Upload README.md with huggingface_hub

Upload benchmark/tabarena/ec2_run.sh with huggingface_hub

wipe corrupt openml cache between phase retries

12x600s phase retries (ride out OpenML outages)

retry phases on transient failure

fleet: spot/restore/checkpoint + systemd self-heal (launch_ec2.py)

fleet: spot/restore/checkpoint + systemd self-heal (ec2_run.sh)

v2: bagged config

Update TabArena EC2 runner scripts (fela_ag_model.py)

Update TabArena EC2 runner scripts (run_tabarena.py)

Update TabArena EC2 runner scripts (ec2_run.sh)

Update TabArena HF job scripts (run_tabarena.py)

Update TabArena HF job script

Update TabArena HF job script

Document TabArena env requirements (separate AutoGluon venv)

Remove stray pycache

Add TabFM/GBM head-to-head benchmark, efficiency tables, and TabArena entry

Update README.md

initial commit