jeffdonahue

#!/bin/bash

INDIR=$1
OUTDIR=$2

echo "Cloning leveldb in ${INDIR} to ${OUTDIR}"

INPATH=`realpath ${INDIR}`

mkdir -p ${OUTDIR}

jeffdonahue

sshput () {
	RSAKEY="${HOME}/.ssh/id_rsa.pub"
	DSAKEY="${HOME}/.ssh/id_dsa.pub"

	if [ $# -eq 2 ]; then
		KEY=$1
		if [ ! -r $KEY ]
		then
			echo "'$KEY' does not exist or is not readable"
			return 1

jeffdonahue

layer {
  name: "data"
  type: "DummyData"
  top: "data"
  top: "label"
  dummy_data_param {
    shape { dim: [128, 3, 227, 227] }
    shape { dim: [128] }
  }
}

jeffdonahue

#!/bin/bash

# Script for installing tmux on systems where you don't have root access.
# tmux will be installed in $HOME/local/bin.
# Either wget or curl is required.
# It's assumed that a C/C++ compiler are installed.

# exit on error
set -e

jeffdonahue

#!/bin/bash

# Script for installing tmux on systems where you don't have root access.
# tmux will be installed in $HOME/local/bin.
# It's assumed that wget and a C/C++ compiler are installed.

# exit on error
set -e

# create our directories

jeffdonahue

#!/usr/bin/env python

# usage: python count_caffemodel_params.py /path/to/my.caffemodel [/path/to/my/other.caffemodel, ...]

from caffe.proto import caffe_pb2
import sys

assert len(sys.argv) >= 2

for caffemodel_filename in sys.argv[1:]:

jeffdonahue

import numpy as np

def canonical_axis(x, axis):
    def is_int_ndarray(x):
        if not isinstance(x, np.ndarray):
            return False
        dtype = str(x.dtype)
        prefixes = 'int', 'uint'
        return any(dtype.startswith(p) for p in prefixes)
    if axis is None:

jeffdonahue

#!/usr/bin/env bash

set -e
set -x

GPU=0
CWD=$(pwd)
CAFFE_DIR=/raid/jdonahue/philkr-caffe
MAGIC_DIR=~/magic_init
CLASS_DIR=/raid/jdonahue/voc-classification

jeffdonahue

#!/bin/bash

set -x
set -e
export LD_LIBRARY_PATH='/usr/local/cuda/lib64/'
export PYTHONUNBUFFERED="True"
NAME="scratch"
NAME="/tmp/magic_caffenet.caffenet"
if [ "$#" -ge 1 ]; then
  NAME=${2}

jeffdonahue

net: "train_val.prototxt"
test_iter: 736
test_interval: 1000000 # py solving tests
display: 1
average_loss: 100
# lr_policy: "fixed"
lr_policy: "step"
stepsize: 50000
gamma: 0.1
# base_lr: 1e-4