Spaces:
Sleeping
Sleeping
Commit History
two-step learning 8af98fa
OutOfMystic commited on
v0.5.1: re-enable decaying height breach penalty from line 4 2038b66
v0.5.0: reduce game_over to -50, disable height breach penalty 3a5b76e
wget modified 8251fe9
OutOfMystic commited on
chore: add version string to game_engine.py (0.4.0) cd64402
feat: decaying height breach penalty based on pieces placed 9dd7378
feat: add one-time height breach penalty (-50 per level above 4) 5a39470
Created using Colab ce596c7
OutOfMystic commited on
Full-game version 3ab55fd
OutOfMystic commited on
fix: reduce step penalty from -1 to -0.1 2ced530
fix: remove height penalty, hole penalty only for new holes 49913c2
Rewrite play.html for per-piece training mode cbba880
Rewrite notebook: per-piece GRPO training on Qwen 3B b61d866
Fix play.html: 100 actions, fix horizontal expansion bug b055f87
Add untrained/trained model demo cells, fix cell ordering 6a04794
Merge branch 'main' of https://github.com/OutOfMystic/tetris-openenv 2b92b06
Created using Colab faa1182
OutOfMystic commited on
Remove broken RUN import check from Dockerfile 41df4cc
Add import check in Dockerfile for debugging 8099afb
Created using Colab fca0f8d
OutOfMystic commited on
Training mode UI + single-char 60-action format d69cfb4
Force HF Space rebuild for /play route 2727176
Add playable Tetris web UI at /play 1d19ab5
Created using Colab e19c278
OutOfMystic commited on
Created using Colab fe74ef3
OutOfMystic commited on
Add random horizontal position to prompts 017d742
Simplify prompts: empty board + random piece from seed 1ea53b8
Fix prompt generation: realistic board states d6a8593
Optimize notebook for A100 GPU 95389d8
Switch to sequence-based training approach ad23073
Fix HF Spaces config: port 8000, base_path /web for OpenEnv web interface fece72b
Enable OpenEnv web interface 37c1d06
Created using Colab 2eae86d
OutOfMystic commited on