import os
import hdl
reload(hdl)
from hdl.models import SparseSlowModel
from hdl.hdl import HDL
from hdl.config import tstring, state_dir
timestring = tstring()
model_base_name = 'HDL_loga_' + timestring + '/layer_%s'
m1 = SparseSlowModel()
layer1_name = 'SparseSlowModel_patchsz020_N512_NN512_l2_subspacel1_dist_2012-05-24_17-33-06/model.model'
fname = os.path.join(state_dir,layer1_name)
m1.load(fname)
m1.model_name = model_base_name % '1'
model_sequence = [
m1,
SparseSlowModel(patch_sz=None, N=512, T=48, sparse_cost='subspacel1', slow_cost='dist', perc_var=99.9, tstring=timestring, model_name=model_base_name % '2'),
SparseSlowModel(patch_sz=None, N=256, T=48, sparse_cost='l1', slow_cost=None, perc_var=99.9, tstring=timestring, model_name=model_base_name % '3')]
hdl_learner = HDL(model_sequence=model_sequence,datasource='PLoS09_Cars_Planes',output_function='proj_loga')
hdl_learner.learn(layer_start=1)
reload(hdl)
from hdl.models import SparseSlowModel
from hdl.hdl import HDL
from hdl.config import tstring
timestring = tstring()
model_base_name = 'HDL_loga_' + timestring + '/layer_%s'
#model_sequence = [
# SparseSlowModel(patch_sz=16, N=200, sparse_cost='subspacel1', slow_cost='dist', tstring=timestring, model_name=model_base_name % '1'),
# SparseSlowModel(patch_sz=None, N=300, sparse_cost='subspacel1', slow_cost='dist', perc_var=99., tstring=timestring, model_name=model_base_name % '2'),
# SparseSlowModel(patch_sz=None, N=100, sparse_cost='l1', slow_cost=None, perc_var=99., tstring=timestring, model_name=model_base_name % '3')]
#model_sequence = [
# SparseSlowModel(patch_sz=32, N=2048, sparse_cost='subspacel1', slow_cost='dist', tstring=timestring, model_name=model_base_name % '1'),
# SparseSlowModel(patch_sz=32, N=2048, sparse_cost='subspacel1', slow_cost='dist', perc_var=99., tstring=timestring, model_name=model_base_name % '2'),
# SparseSlowModel(patch_sz=32, N=2048, sparse_cost='l1', slow_cost=None, perc_var=99., tstring=timestring, model_name=model_base_name % '3')]
#
#hdl_learner = HDL(model_sequence=model_sequence,datasource='vid075-chunks',output_function='infer_loga')
model_sequence = [
SparseSlowModel(patch_sz=48, N=512, T=48, sparse_cost='subspacel1', slow_cost=None, tstring=timestring, model_name=model_base_name % '1'),
SparseSlowModel(patch_sz=None, N=256, T=48, sparse_cost='subspacel1', slow_cost=None, perc_var=99., tstring=timestring, model_name=model_base_name % '2'),
SparseSlowModel(patch_sz=None, N=128, T=48, sparse_cost='l1', slow_cost=None, perc_var=99., tstring=timestring, model_name=model_base_name % '3')]
hdl_learner = HDL(model_sequence=model_sequence,datasource='YouTubeFaces_aligned',output_function='proj_loga')
hdl_learner.learn()
