title: the reward of the first ten steps
source: manuals/miles-101/capture/gpu/out-gb10/read-steps.txt
lines: 1-11
lang: text
note: One line is one step. Read the spread before you read a trend.
---
$ grep -o "'rollout/raw_reward': [0-9.]*" plain.log | head -10
'rollout/raw_reward': 0.62890625
'rollout/raw_reward': 0.59375
'rollout/raw_reward': 0.7578125
'rollout/raw_reward': 0.69140625
'rollout/raw_reward': 0.58984375
'rollout/raw_reward': 0.5859375
'rollout/raw_reward': 0.6875
'rollout/raw_reward': 0.67578125
'rollout/raw_reward': 0.74609375
'rollout/raw_reward': 0.6484375
