vlad17

(gpuc) 14:34:49 vladf@---:/data/vladf$ python tf_random_speed.py 
2017-12-23 14:35:45.782703: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1030] Found device 0 with properties: 
name: Tesla P100-PCIE-16GB major: 6 minor: 0 memoryClockRate(GHz): 1.3285
pciBusID: 0000:0d:00.0
totalMemory: 15.89GiB freeMemory: 14.60GiB
2017-12-23 14:35:45.782759: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1120] Creating TensorFlow device (/device:GPU:0) -> (device: 0, name: Tesla P100-PCIE-16GB, pci bus id: 0000:0d:00.0, compute capability: 6.0)
****************************************
feed_dict (randomize jit) 17.10606360435486 sec
loss 4542.5 -> 4.09381
****************************************

vlad17

(gpuc) 16:32:04 vladf@---:/data/vladf$ python tf_random_speed.py
2017-12-23 16:32:34.773396: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1030] Found device 0 with properties: 
name: Tesla P100-PCIE-16GB major: 6 minor: 0 memoryClockRate(GHz): 1.3285
pciBusID: 0000:0d:00.0
totalMemory: 15.89GiB freeMemory: 14.60GiB
2017-12-23 16:32:34.773443: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1120] Creating TensorFlow device (/device:GPU:0) -> (device: 0, name: Tesla P100-PCIE-16GB, pci bus id: 0000:0d:00.0, compute capability: 6.0)
****************************************
feed_dict (randomize jit) 16.786871671676636 sec
loss 4542.5 -> 4.15892
****************************************

vlad17

# vladimir feinberg 23 dec 2017 tf 1.4
# feel free to use under Apache 2.0 License terms
import time

import numpy as np
import tensorflow as tf

# pretend we don't have X, Y available until we're about
# to train the network, so we have to use placeholders. This is the case
# in, e.g., RL.

vlad17

  accept-process-output(#<process ycmd-server> 0 100 t)
  ycmd--start-server("|\257}\n\362\301\325OW\333&\370I\263\324")
  ycmd-open()
  ycmd--request("/event_notification" (("event_name" . "FileReadyToParse") ("file_data" ("/home/vlad/Documents/cal/courses-year1/cs294-112/cal-deeprl-hw4/controllers.py" ("contents" . "import tensorflow as tf\nimport numpy as np\nimport time\nfrom utils import get_ac_dim, get_ob_dim, build_mlp\n\n\nclass Controller:\n    def __init__(self):\n        pass\n\n    def get_action(self, state):\n        raise NotImplementedError\n\n    def fit(self, data):\n        pass\n\n    def reset(self, nstates):\n        pass\n\nclass RandomController(Controller):\n    def __init__(self, env):\n        super().__init__()\n        self.ac_space = env.action_space\n\n    def get_action(self, states):\n        nstates = len(states)\n        return self._sample_n(nstates)\n\n    def _sample_n(self, n):\n        return np.random.uniform(\n            low=self.ac_space.low,\n            high=self

vlad17

// Use Gists to store code you would like to remember later on
console.log(window); // log the "window" object to the console

vlad17

URLDownload[
  "http://library.wolfram.com/infocenter/MathSource/6770/\
MatrixManipulation.m?file_id=6180", "/tmp/MatrixManipulation.m"];
Quiet@Needs["LinearAlgebra`MatrixManipulation`", 
   "/tmp/MatrixManipulation.m"];
ListLinePlot[
 Table[LinearAlgebra`MatrixManipulation`MatrixConditionNumber[
   RandomReal[
     ]*ToeplitzMatrix[Reverse@Range[n]^10]], {n, 2 ^Range[9]}], 
 PlotLabel -> "poly deg 10"]

vlad17

public class Average implements Sequential<Double, Integer> {
  private long count = 0;
  private long sum = 0;
  private Average() {}
  public static class AverageFactory implements Factory<Average> {
    public Average generate() { return new Average(); }
  }
  public Double apply(Integer x) {
    sum += x;
    count += 1;

vlad17

vlad@vlad-T460:~$ python
Python 3.5.2 (default, Sep 10 2016, 08:21:44) 
[GCC 5.4.0 20160609] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> from scipy.fftpack import dct
>>> import numpy as np
>>> for n in range(1, 5):
...     print(dct(np.identity(n), norm='ortho'))
... 
[[ 1.]]

vlad17

# Subset of the million-song dataset
# Is a song made before or after 2002, based on its vocal features?
# significantly handicapping gbm here:
#  - no cv for stopping
#  - forcing 0.5 threshold.
# The above is what spark has to do

library(gbm)
library(ROCR)

vlad17

// See gbm.R for context
// run with options:
// spark-shell --driver-memory 20G --executor-memory 4G --driver-java-options="-Xss500M" -i gbt.spark
import org.apache.spark.sql.DataFrame
import sys.process._
import java.io._
import org.apache.spark.ml.feature.VectorAssembler
import org.apache.spark.ml.classification.GBTClassifier
import org.apache.spark.ml.evaluation._