Dockerfiles: auto-select torch build (CUDA on x86_64, CPU on arm64); override with TORCH_VARIANT

Scorers: write numeric-only reward.json (Harbor rejects string fields); diagnostics to stdout

README: document self-contained task images (no external base image/repo/token)

Self-contained task images (FROM python:3.12-slim, own requirements.txt); remove external base-image and agent-specific coupling

README: fully agent-neutral (no agent names/specifics), generic -a <agent> run

README: drop AutoDS-specific section; add docker-less (-e singularity) run option for any agent

Reframe README: general benchmark for any Harbor agent; AutoDS specifics optional

README: present Docker (harbor) and venv as two explicit run options

Fix stale README/registry: single dataset, agent-neutral base instructions, data shipped on HF

Add original-competition evaluation details to the 4 full-data MLAB base instructions

Add discretionary GPU hint to base instructions

cifar10 -> full 50k/10k, imdb -> full 25k/25k

Add harbor scaffolding (task.toml/Dockerfile/prepare/tests) to the 4 full-data tasks

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Add files using upload-large-folder tool

Reset mlab-real for 1-folder-per-task restructure

Base instructions from original MLAgentBench repo; fix stale metadata (real data, 9 tasks)

Re-stage identify-contrails on enlarged real subset (~150 records)

Ship real data: spaceship-titanic-base

Ship real data: spaceship-titanic

Ship real data: ogbn-arxiv-base

Ship real data: ogbn-arxiv

Ship real data: imdb-base

Ship real data: imdb

Ship real data: house-price-base

Ship real data: house-price

Ship real data: feedback-base

Ship real data: feedback

Ship real data: cifar10-base

Ship real data: cifar10

Ship real data: amp-parkinsons-base

Ship real data: amp-parkinsons

Ship real vision task data (fathomnet + contrails)

Permissive .gitignore (private mirror ships data)

probe csv in data dir

probe tiny txt

Add real vision task data files (explicit)

Add real vision task data (fathomnet + contrails, private repo)

Add real fathomnet + identify-contrails vision tasks (C1 + baseline)

full dataset incl. real data