from helperFunctions import getUCF101 from helperFunctions import loadFrame import h5py import cv2 from multiprocessing import Pool IMAGE_SIZE = 224 NUM_CLASSES = 101 batch_size = 100 lr = 0.0001 num_of_epochs = 10 data_directory = '/projects/training/bauh/AR/' class_list, train, test = getUCF101(base_directory=data_directory) model = torchvision.models.resnet50(pretrained=True) model.fc = nn.Linear(2048, NUM_CLASSES) for param in model.parameters(): param.requires_grad_(False) for param in model.layer4[2].parameters(): param.requires_grad_(True) for param in model.fc.parameters(): param.requires_grad_(True) params = [] for param in model.layer4[2].parameters():
def main(): data_directory = '/projects/training/bauh/AR/' class_list, train, test = getUCF101(base_directory=data_directory) print_confusion_matrix(class_list=class_list, file_path='part3_confusion_matrix.npy')
import time import math import chainer from chainer import cuda, Function, gradient_check, Variable, optimizers, serializers, utils from chainer import Link, Chain, ChainList import chainer.functions as F import chainer.links as L from helperFunctions import getUCF101 from helperFunctions import getVideoFeatures from helperFunctions import repeatSequence import h5py class_list, train, test = getUCF101() print "LOAD DATA..." for i in xrange(len(train[0])): train[0][i] = train[0][i].split('/')[2].split('.avi')[0] train_hdf5 = h5py.File('train1024.hdf5','r') print "Train..." keys = train_hdf5.keys() t1 = time.time() for key in keys: data = train_hdf5[key][:] t2 = time.time() print t2-t1 for i in xrange(len(test[0])):
from helperFunctions import getUCF101 from helperFunctions import loadFrame import h5py import cv2 from multiprocessing import Pool IMAGE_SIZE = 224 NUM_CLASSES = 101 batch_size = 100 lr = 0.0001 num_of_epochs = 10 data_directory = '/projects/training/bayw/hdf5/' class_list, train, test = getUCF101( base_directory='/projects/training/bayw/hdf5/') model = torchvision.models.resnet50(pretrained=True) model.fc = nn.Linear(2048, NUM_CLASSES) for param in model.parameters(): param.requires_grad_(False) for param in model.layer4[2].parameters(): param.requires_grad_(True) for param in model.fc.parameters(): param.requires_grad_(True) params = [] for param in model.layer4[2].parameters(): params.append(param) for param in model.fc.parameters():