Skip to content

Instantly share code, notes, and snippets.

View coreylynch's full-sized avatar

Corey Lynch coreylynch

  • Google
  • Honolulu, HI
View GitHub Profile
import rclpy
from rclpy.node import Node
import threading
from std_msgs.msg import String
class MinimalPublisher(Node):
def __init__(self, name):
Async 1-step DQN
clone https://github.com/coreylynch/async-rl
run python async_dqn.py --experiment breakout --game "Breakout-v0" --num_concurrent 8
@coreylynch
coreylynch / gist:c294496e7c42946798c67e7f5ba94259
Last active June 18, 2017 09:50
Setting up tensorflow / keras / gym on an c4.8xlarge EC2 instance
# Not a bash script, just the commands that worked for me
# C compiler
sudo yum groupinstall 'Development Tools'
# Virtualenv
sudo pip install --upgrade virtualenv
virtualenv --system-site-packages ~/tensorflow
# Tensorflow (virtualenv)
Title: Walk in the blizzard
With some IFs, I am thinking of a walk in the forecasted blizzard on Saturday. My departure time will be about 10:30 AM and I would expect to be walking at a brisk, but not fast pace. My maximum destination will be Fairway in Red Hook with a stop there for lunch, something hot to drink and a bathroom stop. Walking with no snow is one hour and I would guess with the snow and blizzard conditions it would be more like an hour and a half. Roundtrip with the stop could be as much as four hours.
Here are the IFs:
If there is no rain and it is all snow.
If Fairway is open.
If the snow is not too extreme.
For sure I'll spend some time trudging through the storm.
In [21]: fm = pylibfm.FM(num_factors=10, num_iter=600, verbose=True, task="regression", initial_learning_rate=0.0001, learning_rate_schedule=
"optimal")
In [22]: fm.fit(X_train,y_train)
Creating validation dataset of 0.01 of training for adaptive regularization
-- Epoch 1
Training RMSE: 0.78393
-- Epoch 2
Training RMSE: 0.61129
In [19]: fm = pylibfm.FM(num_factors=10, num_iter=300, verbose=True, task="regression", initial_learning_rate=0.0001, learning_rate_schedule=
"optimal")
In [20]: fm.fit(X_train,y_train)
Creating validation dataset of 0.01 of training for adaptive regularization
-- Epoch 1
Training RMSE: 0.78393
-- Epoch 2
Training RMSE: 0.61129
-- Epoch 3
Listen to squarespace tryna get in on this
Me: Oh cool a credible witness debunking the best buy phone booth theory.
Her: “I used to steal cds from there all the time”.
Me: fuuuck
2:36 not possible b/c she was with this wrestling girl Summer at the time.
Not the only person who saw her at 3pm
with open("predictions.txt","wb") as f:
for i in range(len(test_data)):
line = test_data[i]["user_id"]+"\t"+test_data[i]["movie_id"]+"\t"+str(preds[i])
f.write(line+"\n")
*##X
*###*
####X
*# ####X#
X#X #X##X#
X *## ###X#X
X# *XX* *##X X####
##* *######XX# # *##X
### *#X #X* # ** ##
#X###X X ## ***#### ##*
20110908_NO@GB,1,,0,NO,GB,,,,T.Morstead kicks 68 yards from NO 35 to GB -3. R.Cobb to GB 24 for 27 yards (L.Torrence).,0,0,2011
20110908_NO@GB,1,59,55,GB,NO,1,10,76,(14:55) (Shotgun) A.Rodgers pass short right to G.Jennings to GB 33 for 9 yards (T.Porter).,0,0,2011
20110908_NO@GB,1,59,25,GB,NO,2,1,67,(14:25) (Shotgun) R.Grant right tackle to GB 37 for 4 yards (C.Jordan).,0,0,2011
20110908_NO@GB,1,58,58,GB,NO,1,10,63,(13:58) A.Rodgers pass short right to J.Nelson to GB 43 for 6 yards (J.Greer).,0,0,2011
20110908_NO@GB,1,58,25,GB,NO,2,4,57,(13:25) (Shotgun) A.Rodgers sacked at GB 35 for -8 yards (J.Casillas).,0,0,2011
20110908_NO@GB,1,57,51,GB,NO,3,12,65,(12:51) (Shotgun) A.Rodgers pass short right to D.Driver to NO 49 for 16 yards (M.Jenkins).,0,0,2011
20110908_NO@GB,1,57,17,GB,NO,1,10,49,(12:17) (Shotgun) A.Rodgers pass deep left to J.Nelson to NO 13 for 36 yards (P.Robinson).,0,0,2011
20110908_NO@GB,1,56,29,GB,NO,1,10,13,(11:29) (Shotgun) A.Rodgers scrambles up the middle to NO 12 for 1 yard (J.Casillas).,0,