Compact AGENTS.md into zero-memory survival guide

Upload phase2_ckpt_650000.zip with huggingface_hub

Add instruction: always update docs before starting long-running tasks

Update AGENTS.md: document how hf_jobs script parameter actually works (converts to raw Hub URL)

Upload phase3_curriculum.py

Upload session_state.json

Upload phase2_resume.py

Upload smoke_test_ckpt.zip with huggingface_hub

Upload smoke_test_v2.py

Add HF Jobs research — snapshot_download for private repos, script submission pattern

Add smoke_test.py for HF Job validation

Add Sandbox Policy section per user mandate

Add AGENTS.md — persistent context and lessons learned

Add session_state.json for cross-session persistence

Add Phase 2 HF Job training script

Upload phase2_ckpt_600352.zip with huggingface_hub

Upload phase2_ckpt_550352.zip with huggingface_hub

Update docs: Phase 2 started, add interim results

Add comprehensive documentation for TIL-26-AE project

Upload phase1_summary.txt with huggingface_hub

Upload phase1_eval_results.txt with huggingface_hub

Upload phase1_final.zip with huggingface_hub

Upload ckpt_400000.zip with huggingface_hub

Upload ckpt_350000.zip with huggingface_hub

Upload ckpt_300000.zip with huggingface_hub

Upload ckpt_250000.zip with huggingface_hub

Add inference-ready AE manager for loading trained MaskablePPO

Upload ckpt_200000.zip with huggingface_hub

Upload ckpt_150000.zip with huggingface_hub

Upload ckpt_100000.zip with huggingface_hub

Upload ckpt_50000.zip with huggingface_hub

Add train_in_space.py for running training inside the Space

Fix agent tracking to use possible_agents instead of agents attribute

Upload train_all_phases.py

Upload train_all_phases.py

Update ML Intern artifact metadata

Upload requirements.txt

Upload README.md

Upload train_all_phases.py

Update ML Intern artifact metadata

initial commit