mlzxy

SAFE-RM

It behaves identically asrm except that it won't delete files with important or .keep (customizable through SAFE_RM_IMPORTANCE_MARK ). The idea is to mark important files beforehand so they can always survive future disk cleanup.

What triggered me to create this is I accidentally removed some baseline models, which I spent lots of GPU hours to train and all my current training jobs depend on. To prevent future disasters, I come up with this solution. Hope you find it useful too.

Demo

mlzxy

# in real-world scenaria, computational resource is usually fixed
# therefore, we only use time budget here
game_tree = Node(current_state)
while time_budget > 0:
  start_time = time()
  build(game_tree)
  time_budget -= time() - start_time

mlzxy

# in cfr, we simulate all actions
v[I] = {a: cfr(h + [a], {**π_i, P(h): π_i[P(h)] * σ[t][I][a]}, i, t) 
    						  for a in A[I]} 

# in outcome sampling mccfr, we only need to sample one a from A[I]
a = sample(A[I], σ[t][I]) # or use `ϵ * uniform + (1-ϵ) * σ[t][I]`
v[I][a] = mccfr(h + [a], {**π_i, P(h): π_i[P(h)] * σ[t][I][a]})

mlzxy

def cfr(
    h=[],    # history
    π_i={},  # a mapping from player_id to π_i(h)
    i=0,     # player_id 
    t=0      # timestep
): 
    # here I skip the chance node case for simplicity
    if h is terminal:
    	z = h
    	return u(i, z)  # utility u(z) for player i

mlzxy

#!/bin/bash
# hook into inotify to watch a file, and generate
# diffs for changes in real-time
#
# requires inotify-tools (apt-get install inotify-tools)
#
# @author Filipe Dobreira <filipe.dobreira@ez.no>
usage() {
  echo "Usage:"
  echo "  in-diff <file>|help"

mlzxy

import argparse
import logging
import random
import time
import mxnet as mx
from mxnet import nd
from mxnet import image
from mxnet import gluon
from mxnet import autograd
import numpy as np

mlzxy

import argparse, time
import logging
logging.basicConfig(level=logging.INFO)
fh = logging.FileHandler('training.log')
logger = logging.getLogger()
logger.addHandler(fh)

import mxnet as mx
from mxnet import gluon
from mxnet.gluon import nn

mlzxy

array=("human-computer-interaction" "design-principles" "social-computing" "interaction-techniques" "user-research" "infodesign" "designexperiments" "interaction-design-capstone")

for i in ${array[@]}; do
    ../coursera-dl -n ~/.netrc  ${i}
done

echo "download finished"

mlzxy

from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

import tensorflow as tf


class GRU(tf.contrib.rnn.RNNCell):

  def __init__(

mlzxy

name: "Inception_Resnet2_Imagenet"
layer {
  name: "data"
  type: "Data"
  top: "data"
  top: "label"
  include {
    phase: TRAIN
  }
  transform_param {