def pred_loop():
# evalu_model, x_ = eval_model.build()
global pred_file_name
evalu_model, x_ = eval_model.build()
while True:
# print "waiting"
# s_time = time.time()
print "pred_loop waiting"
vid, cur_fr,n,fileName = q.get()
# print "waited",(time.time()-s_time)/1000.,"ms"
pred_file_name = fileName.split('/')
pred_file_name = pred_file_name[-1].replace(".zip","_prediction.csv")
# print "get"
# print vid.shape, cur_fr
x_.set_value(vid.astype("float32"),borrow=True)
pred = evalu_model()[0][0]
pred_p = pred.max()
pred_idx = pred.argmax()+1
# fps = int(1./((time.time()-time_start)/step))
fr_start = cur_fr+step-n_f
fr_end = cur_fr+step
# print pred_idx,"\t", "\t",pred_p,"\t",fr_start, "-",fr_end,"\t",fps,'fps'
predict(pred_idx,pred_p,fr_start)
# print v_new.shape
# for i in xrange(v_new.shape[0]):
# for j in xrange(v_new.shape[1]):
# for k in xrange(v_new.shape[2]):
# play_vid(v_new[i,j,k],wait=0)
# cur_fr += step
# print cur_fr, int(1./((time.time()-time_start)/step)),'fps'
if cur_fr+2*step>=n:
reinit()
# data = "/media/Data/mp/chalearn2014/40x90x90_train"
# data = "/home/lio/mp/chalearn2014/train_raw"
# data = "/home/lio/mp/chalearn2014/valid"
# dst = "/home/lio/Dropbox/MP/chalearn2014/evaluation/results/test"
data = "E:/mp/chalearn2014/train_raw"
dst = "C:/Users/lio/Dropbox/MP/chalearn2014/evaluation/results/test"
pp.make_sure_path_exists(dst)
step = 5
n_f = 32
pred_file_name = ''
h = vid_shape[-1]
# q = Queue(10)
# q2 = Queue(20)
evalu_model, x_ = eval_model.build()
# def eval(fileName):
# @profile
def eval():
global pred_file_name
# pred_file_name = fileName.split('/')
# pred_file_name = pred_file_name[-1].replace(".zip","_prediction.csv")
# print fileName
files = glob(data+"/"+"*.zip")
files.sort()
print len(files), "found"
