arthur/handson-ml3
1
0
Fork 0
mirror of https://github.com/ArthurDanjou/handson-ml3.git synced 2026-01-14 20:19:29 +01:00
Code Issues Packages Projects Releases Wiki Activity

use list with circular indexing instead of deque as the replay buffer

Browse Source
This commit is contained in:
Nilesh PS
2018-03-28 19:57:26 +05:30
parent 9ac3c27e7b
commit 16e8a8cf61
1 changed files with 17 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

17
16_reinforcement_learning.ipynb
View File

@@ -1576,6 +1576,23 @@
"replay_memory_size = 500000\n",
"replay_memory = deque([], maxlen=replay_memory_size)\n",
"\n",
"\n",
"class ReplayMemory:\n",
" \n",
" def __init__(self, size):\n",
" self.size = size\n",
" self.buf= [None] * self.size\n",
" self.index = 0\n",
" \n",
" def append(self, data):\n",
" self.buf[self.index] = data\n",
" self.index = (self.index + 1) % self.size\n",
" \n",
" def __getitem__(self, idx):\n",
" return self.buf[idx]\n",
"\n",
" \n",
"\n",
"def sample_memories(batch_size):\n",
" indices = np.random.permutation(len(replay_memory))[:batch_size]\n",
" cols = [[], [], [], [], []] # state, action, reward, next_state, continue\n",