issues
search
DeNA
/
HandyRL
HandyRL is a handy and simple framework based on Python and PyTorch for distributed reinforcement learning that is applicable to your own environments.
MIT License
282
stars
42
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Monte Carlo Tree Search :: AlphaGeese, AlphaZero .. etc
#356
Chandanpanda
opened
2 months ago
0
(2024/04) merge develop into master
#355
YuriCat
closed
6 months ago
0
(Outputted results will be changed) output whole results
#354
YuriCat
opened
7 months ago
0
fix: integer value for randrange()
#353
YuriCat
closed
8 months ago
0
(2024/01) merge develop into master
#352
YuriCat
closed
9 months ago
0
fix: value padding with [0]
#351
YuriCat
closed
9 months ago
0
fix: disconnect after socket.timeout
#350
YuriCat
closed
7 months ago
0
Revert "chore: remove fileno() inferface from PickledConnection"
#349
YuriCat
closed
12 months ago
1
(2023/11) merge develop into master
#348
YuriCat
closed
11 months ago
0
feature: remove Python 3.7 from github CI actions
#347
YuriCat
closed
1 year ago
0
Replay Buffer
#346
Chandanpanda
opened
1 year ago
2
fix: fill 0 for reward, return, value in make_batch()
#345
YuriCat
closed
1 year ago
0
feature: remove installation of kaggle environments from github actions
#344
YuriCat
closed
1 year ago
0
Large scale training
#343
Jogima-cyber
closed
1 year ago
3
feature: model pool in each worker
#342
YuriCat
opened
1 year ago
0
feature: data gathering code both for solo/multi player training
#341
YuriCat
closed
1 year ago
1
feature: apply omask for two-player value averaging for solo-play
#340
YuriCat
closed
1 year ago
1
fix: remove map_location=cpu error after starting training
#339
YuriCat
closed
1 year ago
1
feature: return dict from evaluation function (same key)
#338
YuriCat
closed
1 year ago
0
feature: Specify opponent by path
#337
sakami0000
opened
1 year ago
0
fix: prevent array length error in win_rate_plot.py
#336
YuriCat
closed
1 year ago
0
feature: remove eval_coef from config.yaml and set 0.85 as default value
#335
YuriCat
opened
1 year ago
0
(2022/11-12) merge develop into master
#334
YuriCat
closed
1 year ago
0
feature: remove resolve_agent and use build_agent to build trained model agent
#333
YuriCat
opened
1 year ago
0
feature: remove ModelWrapper from train.py
#332
YuriCat
opened
1 year ago
0
fix: episode selection index error
#331
YuriCat
closed
1 year ago
0
feature: combine worker argument both from trainer and worker
#330
YuriCat
opened
1 year ago
0
feature: eval_rate -> eval_coef
#329
YuriCat
opened
1 year ago
0
Adding multidiscrete feature
#328
Jogima-cyber
opened
2 years ago
9
feature: accept any structure of results
#327
YuriCat
closed
2 years ago
1
feature: add python3.10 test
#326
YuriCat
closed
1 year ago
0
Merge develop branch into master, August 2022.
#325
ikki407
closed
2 years ago
0
(Idea) feature: proportional accept rate during all phases
#324
YuriCat
closed
2 years ago
1
(IMPORTANT) Deep Nash algorithm
#323
YuriCat
opened
2 years ago
0
feature: compute rho, c by joint probability
#322
YuriCat
opened
2 years ago
0
(Change Default Outputs) feature: change default learning rate to 3e-6 * sqrt(batch_size)
#321
YuriCat
opened
2 years ago
0
chore: add kwargs to random model
#320
YuriCat
closed
2 years ago
1
(2022/05) merge develop into master
#319
ikki407
closed
2 years ago
0
fix: stop calling view_transition() in reset phase
#318
YuriCat
closed
2 years ago
0
feature: divide ep count variable
#317
YuriCat
closed
2 years ago
1
feature: set show flag in game_args
#316
YuriCat
opened
2 years ago
0
feature: opponent selection by :
#315
YuriCat
closed
2 years ago
1
(WIP) google research football environment
#314
YuriCat
opened
2 years ago
0
fix: set key for rule-based agents
#313
YuriCat
closed
2 years ago
0
feature: rulebase agents with keys
#312
YuriCat
closed
2 years ago
1
feature: connection count interface
#311
YuriCat
closed
2 years ago
1
feature: set model=None in load_model()
#310
YuriCat
closed
2 years ago
0
feature: use set instead of list for connections in QueueCommunicator
#309
YuriCat
closed
2 years ago
1
(2022/03-04) merge develop into master
#308
ikki407
closed
2 years ago
0
(Idea) feature: daemonic threads & daemonic batcher processes & wait stopping workers
#307
YuriCat
closed
2 years ago
1
Next