chicm

import os
import cv2
import collections
import time 
import tqdm
from PIL import Image
from functools import partial
train_on_gpu = True

import numpy as np

chicm

""" 
Ensembling methods for object detection.
"""

""" 
General Ensemble - find overlapping boxes of the same class and average their positions
while adding their confidences. Can weigh different detectors with different weights.
No real learning here, although the weights and iou_thresh can be optimized.

Input: 

chicm

with open('h5c/h5ctrain.zh', 'r') as f:
    text = f.readlines()
    text = [line.strip() for line in text]
print(text[:10])

split_text = []
for line in text:
    newline = ''
    for c in line:

chicm

import glob,os


def load_properties(props, filepath, sep='=', comment_char='#'):
    """
    Read the file passed as parameter as a properties file.
    """
    if not os.path.exists(filepath):
        return
    with open(filepath, "rt") as f:

chicm

"""
This is a batched LSTM forward and backward pass
"""
import numpy as np
import code

class LSTM:
  
  @staticmethod
  def init(input_size, hidden_size, fancy_forget_bias_init = 3):

chicm

FROM: http://benanne.github.io/2015/03/17/plankton.html

Meta-Tricks:
	- Use %10 for validation with STRATIFIED SAMPLING (my mistake)
	- Cyclic Pooling 
	- Leaky ReLU = max(x, a*x) learned a
		- reduces overfitting with a ~= 1/3
	- Orthogonal initialization http://arxiv.org/pdf/1312.6120v3.pdf
	- Use larger weight decay for larger models since otherwise some layers might diverge

chicm

FROM: http://benanne.github.io/2015/03/17/plankton.html

Meta-Tricks:
	- Use %10 for validation with STRATIFIED SAMPLING (my mistake)
	- Cyclic Pooling 
	- Leaky ReLU = max(x, a*x) learned a
		- reduces overfitting with a ~= 1/3
	- Orthogonal initialization http://arxiv.org/pdf/1312.6120v3.pdf
	- Use larger weight decay for larger models since otherwise some layers might diverge

chicm

'''This script goes along the blog post
"Building powerful image classification models using very little data"
from blog.keras.io.
It uses data that can be downloaded at:
https://www.kaggle.com/c/dogs-vs-cats/data
In our setup, we:
- created a data/ folder
- created train/ and validation/ subfolders inside data/
- created cats/ and dogs/ subfolders inside train/ and validation/
- put the cat pictures index 0-999 in data/train/cats

chicm

##VGG16 model for Keras

This is the Keras model of the 16-layer network used by the VGG team in the ILSVRC-2014 competition.

It has been obtained by directly converting the Caffe model provived by the authors.

Details about the network architecture can be found in the following arXiv paper:

Very Deep Convolutional Networks for Large-Scale Image Recognition

K. Simonyan, A. Zisserman

chicm

#include <bits/stdc++.h>
#define abs(x) ((x) < 0?-(x):(x))
#define uint unsigned int
using namespace std;

int gcd(int x, int y) {
   if(x > y) swap(x,y);
   while(x > 0) {
      int z =x;
      x =y%x;