ml-agents

目录树: c42eff57

作者	SHA1	备注	提交日期
Arthur Juliani	de700c3a	Multi Brain Training and Recurrent state encoder (#166 ) * `learn.py` is now main script for training brains. * Simultaneous multi-brain training is now possible. * `ghost-trainer` allows for proper training in adversarial scenarios. * `imitation-trainer` provides a basic implementation of real-time behavioral cloning. * All trainer hyperparameters now exist in `.yaml` files. * `PPO.ipynb` removed. * LSTM model added. * More dynamic buffer class to handle greater variety of scenarios.	7 年前
vincentpierre	a54e459c	partial fix on the lstm The recurrent encoding now happens at the end	7 年前
GitHub	51621334	State Stacking & Banan Environment (#262 ) * Add support for stacking past n states to allow network to learn temporal dependencies. * Add Banana Collector environment for demonstrating partially observable multi-agent environments. * Add 3DBall Hard which lacks velocity information in state representation. Used as test for LSTM and state-stacking features. * Rework Tennis environment to be continuous control and trainable in 100k steps.	7 年前
GitHub	f8a8b112	Move epsilon generation into graph (#283 )	7 年前
GitHub	36d58cee	Add Seeding, MaxStepReached, and Bootstrapping fix (#303 ) * Add ability to seed learning (numpy, tensorflow, and Unity) with `--seed` flag. * Add `maxStepReached` flag to Agents and Academy. * Change way value bootstrapping works in PPO to take advantage of timeouts. * Default size of GridWorld changed to 5x5 in order to validate bootstrapping changes.	7 年前
GitHub	e676017b	Reorganize learn.py (#302 ) Split learn.py into learn.py as command-line wrapper, and trainer_controller.py as core trainer/env logic.	7 年前
Arthur Juliani	6ad7f010	Fix for discrete control image observations	7 年前
Arthur Juliani	c42eff57	Misc fixes	7 年前

8 次代码提交 (c42eff57-de41-40a1-a69d-e7dfdc556646)