Made inference from the saved model (d8e8db9f) · Commits · Elphel / python3-imagej-tiff

infer_qcds_01.py

+39 −36

Original line number	Original line	Diff line number	Diff line
	@@ -4,15 +4,14 @@ __copyright__ = "Copyright 2018, Elphel, Inc."
	__license__ = "GPL-3.0+"		__license__ = "GPL-3.0+"
	__email__ = "andrey@elphel.com"		__email__ = "andrey@elphel.com"

	#python3 nn_ds_neibs17.py /home/eyesis/x3d_data/data_sets/conf/qcstereo_conf13.xml /home/eyesis/x3d_data/data_sets		#Builds (and saved) inference model from trained by nn_ds_neibs21.py
			#Model and weights are used by the inference-only infer_qcds_graph.py
	import os		import os
	import sys		import sys
	import numpy as np		import numpy as np
	import time		import time
	import shutil		import shutil
	from threading import Thread
	import qcstereo_network		import qcstereo_network
	import qcstereo_losses
	import qcstereo_functions as qsf		import qcstereo_functions as qsf
	import tensorflow as tf		import tensorflow as tf
	from tensorflow.python.ops import resource_variable_ops		from tensorflow.python.ops import resource_variable_ops
	@@ -120,26 +119,10 @@ image_data = qsf.initImageData( # just use image_data[0]
	infer = True,		infer = True,
	keep_gt = True) # to generate same output files		keep_gt = True) # to generate same output files

	ph_corr2d = tf.placeholder(np.float32, (None,FEATURES_PER_TILE))		ph_corr2d = tf.placeholder(np.float32, (None,FEATURES_PER_TILE), name = 'ph_corr2d')
	ph_target_disparity = tf.placeholder(np.float32, (None,1))		ph_target_disparity = tf.placeholder(np.float32, (None,1), name = 'ph_target_disparity')
	#ph_tile_yx = tf.placeholder(np.int32, (None,2)) #tile_y, tile_x		ph_ntile = tf.placeholder(np.int32, (None,), name = 'ph_ntile') #nTile
	ph_ntile = tf.placeholder(np.int32, (None,)) #nTile		ph_ntile_out = tf.placeholder(np.int32, (None,), name = 'ph_ntile_out') #which tiles should be calculated in stage2
	ph_ntile_out = tf.placeholder(np.int32, (None,)) #which tiles should be calculated in stage2



	"""
	dataset_img = tf.data.Dataset.from_tensor_slices({
	"corr2d": corr2d_placeholder,
	"target_disparity": target_disparity_placeholder,
	"xy": tile_xy_placeholder})
	dataset_img = dataset_img.prefetch(image_data[0]['corr2d'].shape[0])
	iterator_img = dataset_img.make_initializable_iterator()
	next_element_tt = iterator_img.get_next()
	"""
	#No need to use datasets here = whole input even for the full frame( ~100MB) should fit



	#corr2d9x325 = tf.concat([tf.reshape(next_element_tt['corr2d'],[-1,cluster_size,FEATURES_PER_TILE]) , tf.reshape(next_element_tt['target_disparity'], [-1,cluster_size, 1])],2)		#corr2d9x325 = tf.concat([tf.reshape(next_element_tt['corr2d'],[-1,cluster_size,FEATURES_PER_TILE]) , tf.reshape(next_element_tt['target_disparity'], [-1,cluster_size, 1])],2)
	tf_intile325 = tf.concat([ph_corr2d, ph_target_disparity],axis=1,name="tf_intile325") # [?,325]		tf_intile325 = tf.concat([ph_corr2d, ph_target_disparity],axis=1,name="tf_intile325") # [?,325]
	@@ -156,7 +139,8 @@ Probably ResourceVariable is not needed here because of the tf.scatter_update()
	If collection is not provided, it defaults to [GraphKeys.GLOBAL_VARIABLES], and that in turn fails saver.restore() as this variable was not available in the trained model		If collection is not provided, it defaults to [GraphKeys.GLOBAL_VARIABLES], and that in turn fails saver.restore() as this variable was not available in the trained model
	"""		"""

	rv_stage1_out = resource_variable_ops.ResourceVariable(		#rv_stage1_out = resource_variable_ops.ResourceVariable(
			rv_stage1_out = tf.Variable(
	np.zeros([HEIGHT * WIDTH, NN_LAYOUT1[-1]]),		np.zeros([HEIGHT * WIDTH, NN_LAYOUT1[-1]]),
	## collections = [],		## collections = [],
	collections = [GraphKeys.LOCAL_VARIABLES],# Works, available with tf.local_variables()		collections = [GraphKeys.LOCAL_VARIABLES],# Works, available with tf.local_variables()
	@@ -188,7 +172,7 @@ tf_stage2_in_sparse = tf.gather(tf_stage2_in, indices= ph_ntile_out, axis=0, nam
	#ext=np.concatenate([aextv[:1,:,:]]1 + [aextv] + [aextv[-1:,:,:]]3,axis = 0)		#ext=np.concatenate([aextv[:1,:,:]]1 + [aextv] + [aextv[-1:,:,:]]3,axis = 0)


	with tf.name_scope("Siam_net"): # to have the same scope for weight/biases?		with tf.name_scope("Disparity_net"): # to have the same scope for weight/biases?
	ns, _ = qcstereo_network.network_sub(tf_intile325,		ns, _ = qcstereo_network.network_sub(tf_intile325,
	input_global = [None,ph_target_disparity][SPREAD_CONVERGENCE], # input_global[:,i,:],		input_global = [None,ph_target_disparity][SPREAD_CONVERGENCE], # input_global[:,i,:],
	layout= NN_LAYOUT1,		layout= NN_LAYOUT1,
	@@ -203,24 +187,41 @@ with tf.name_scope("Siam_net"): # to have the same scope for weight/biases?
	with tf.control_dependencies([update]):		with tf.control_dependencies([update]):
	stage1done = tf.constant(1, dtype=tf.int32, name="stage1done")		stage1done = tf.constant(1, dtype=tf.int32, name="stage1done")
	pass		pass
	stage2_out_sparse = qcstereo_network.network_inter (		stage2_out_sparse0 = qcstereo_network.network_inter (
	input_tensor = tf_stage2_in_sparse,		input_tensor = tf_stage2_in_sparse,
	input_global = None, # [None, ig][inter_convergence], # optionally feed all convergence values (from each tile of a cluster)		input_global = None, # [None, ig][inter_convergence], # optionally feed all convergence values (from each tile of a cluster)
	layout = NN_LAYOUT2,		layout = NN_LAYOUT2,
	reuse = False,		reuse = False,
	use_confidence = False)		use_confidence = False)
			stage2_out_sparse = tf.identity(stage2_out_sparse0, name = 'stage2_out_sparse')

	if not USE_SPARSE_ONLY: #Does it reduce the graph size?		if not USE_SPARSE_ONLY: #Does it reduce the graph size?
	stage2_out_full = qcstereo_network.network_inter (		stage2_out_full0 = qcstereo_network.network_inter (
	input_tensor = tf_stage2_in,		input_tensor = tf_stage2_in,
	input_global = None, # [None, ig][inter_convergence], # optionally feed all convergence values (from each tile of a cluster)		input_global = None, # [None, ig][inter_convergence], # optionally feed all convergence values (from each tile of a cluster)
	layout = NN_LAYOUT2,		layout = NN_LAYOUT2,
	reuse = True,		reuse = True,
	use_confidence = False)		use_confidence = False)
			stage2_out_full = tf.identity(stage2_out_full0, name = 'stage2_out_full')

	pass		pass

	ROOT_PATH = './attic/infer_qcds_graph'+SUFFIX+"/" # for tensorboard		ROOT_PATH = './attic/infer_qcds_graph'+SUFFIX+"/" # for tensorboard
			"""
	saver=tf.train.Saver()		This is needed if ResourceVariable is used - then i/o tensors names somehow disappeared
			and were replaced by 'Placeholder_*'
			collection_io = 'collection_io'
			tf.add_to_collection(collection_io, ph_corr2d)
			tf.add_to_collection(collection_io, ph_target_disparity)
			tf.add_to_collection(collection_io, ph_ntile)
			tf.add_to_collection(collection_io, ph_ntile_out)
			tf.add_to_collection(collection_io, stage1done)
			tf.add_to_collection(collection_io, stage2_out_sparse)
			"""
			##saver=tf.train.Saver()
			saver=tf.train.Saver(tf.global_variables())
			saver_def = saver.as_saver_def()
			pass
	"""		"""
	saver_def = saver.as_saver_def()		saver_def = saver.as_saver_def()

	@@ -238,11 +239,18 @@ saver_def.restore_op_name= save/restore_all
	saver_def.save_tensor_name= save/control_dependency:0		saver_def.save_tensor_name= save/control_dependency:0
	print(saver.save(sess, files["checkpoints"]))		print(saver.save(sess, files["checkpoints"]))
	"""		"""
			try:
			os.makedirs(os.path.dirname(files['inference']))
			print ("Created directory ",os.path.dirname(files['inference']))
			except:
			pass


	with tf.Session() as sess:		with tf.Session() as sess:
	sess.run(tf.global_variables_initializer())		sess.run(tf.global_variables_initializer())
	sess.run(tf.local_variables_initializer())		sess.run(tf.local_variables_initializer())
	saver.restore(sess, files["checkpoints"])		saver.restore(sess, files["checkpoints"])
			saver.save(sess, files["inference"]) #TODO: move to different subdir
	merged = tf.summary.merge_all()		merged = tf.summary.merge_all()
	writer = tf.summary.FileWriter(ROOT_PATH, sess.graph)		writer = tf.summary.FileWriter(ROOT_PATH, sess.graph)
	lf = None		lf = None
	@@ -288,12 +296,6 @@ with tf.Session() as sess:
	[disp_out.reshape(-1,1),		[disp_out.reshape(-1,1),
	dataset_img['t_disps'], #t_disps[ntest],		dataset_img['t_disps'], #t_disps[ntest],
	dataset_img['gtruths'], # gtruths[ntest],		dataset_img['gtruths'], # gtruths[ntest],
	# dbg_cost_nw.reshape(-1,1),
	# dbg_cost_w.reshape(-1,1),
	# dbg_d.reshape(-1,1),
	# dbg_avg_disparity.reshape(-1,1),
	# dbg_gt_disparity.reshape(-1,1),
	# dbg_offs.reshape(-1,1)
	],1)		],1)
	np.save(result_file, rslt.reshape(HEIGHT,WIDTH,-1))		np.save(result_file, rslt.reshape(HEIGHT,WIDTH,-1))
	rslt = qsf.eval_results(result_file, ABSOLUTE_DISPARITY, radius=CLUSTER_RADIUS, logfile=lf) # (re-loads results). Only uses first 4 layers		rslt = qsf.eval_results(result_file, ABSOLUTE_DISPARITY, radius=CLUSTER_RADIUS, logfile=lf) # (re-loads results). Only uses first 4 layers
	@@ -305,6 +307,7 @@ with tf.Session() as sess:
	"""		"""
	image_data[nimg] = None		image_data[nimg] = None

			meta_graph_def = tf.train.export_meta_graph(files["inference"]+'.meta')


	if lf:		if lf:

infer_qcds_graph_01.py

0 → 100644

+196 −0

Original line number	Original line	Diff line number	Diff line
			#!/usr/bin/env python3
			__copyright__ = "Copyright 2018, Elphel, Inc."
			__license__ = "GPL-3.0+"
			__email__ = "andrey@elphel.com"

			# Just inference, currently uses /data_ssd/data_sets/tf_data_5x5_main_13_heur/inference/
			import os
			import sys
			import numpy as np
			import time
			import shutil
			##import qcstereo_network
			import qcstereo_functions as qsf
			import tensorflow as tf
			#from tensorflow.python.ops import resource_variable_ops
			#tf.ResourceVariable = resource_variable_ops.ResourceVariable

			qsf.TIME_START = time.time()
			qsf.TIME_LAST = qsf.TIME_START

			IMG_WIDTH = 324 # tiles per image row
			DEBUG_LEVEL= 1

			try:
			conf_file = sys.argv[1]
			except IndexError:
			print("Configuration path is required as a first argument. Optional second argument specifies root directory for data files")
			exit(1)
			try:
			root_dir = sys.argv[2]
			except IndexError:
			root_dir = os.path.dirname(conf_file)

			print ("Configuration file: " + conf_file)
			parameters, dirs, files, _ = qsf.parseXmlConfig(conf_file, root_dir)
			"""
			Temporarily for backward compatibility
			"""
			if not "SLOSS_CLIP" in parameters:
			parameters['SLOSS_CLIP'] = 0.5
			print ("Old config, setting SLOSS_CLIP=", parameters['SLOSS_CLIP'])

			"""
			Defined in config file
			"""
			TILE_SIDE, TILE_LAYERS, TWO_TRAINS, NET_ARCH1, NET_ARCH2 = [None]*5
			ABSOLUTE_DISPARITY,SYM8_SUB, WLOSS_LAMBDA, SLOSS_LAMBDA, SLOSS_CLIP = [None]*5
			SPREAD_CONVERGENCE, INTER_CONVERGENCE, HOR_FLIP, DISP_DIFF_CAP, DISP_DIFF_SLOPE = [None]*5
			CLUSTER_RADIUS = None
			PARTIALS_WEIGHTS, MAX_IMGS_IN_MEM, MAX_FILES_PER_GROUP, BATCH_WEIGHTS, ONLY_TILE = [None] * 5
			USE_CONFIDENCE, WBORDERS_ZERO, EPOCHS_TO_RUN, FILE_UPDATE_EPOCHS = [None] * 4
			LR600,LR400,LR200,LR100,LR = [None]*5
			SHUFFLE_FILES, EPOCHS_FULL_TEST, SAVE_TIFFS = [None] * 3
			CHECKPOINT_PERIOD = None
			TRAIN_BUFFER_GPU, TRAIN_BUFFER_CPU = [None]*2
			TEST_TITLES = None
			USE_SPARSE_ONLY = True
			LOGFILE="results-infer.txt"
			"""
			Next gets globals from the config file
			"""
			globals().update(parameters)

			WIDTH = 324
			HEIGHT = 242
			TILE_SIZE = TILE_SIDE* TILE_SIDE # == 81
			FEATURES_PER_TILE = TILE_LAYERS * TILE_SIZE# == 324
			BATCH_SIZE = ([1,2][TWO_TRAINS])21000//25 # == 80 Each batch of tiles has balanced D/S tiles, shuffled batches but not inside batches

			SUFFIX=(str(NET_ARCH1)+'-'+str(NET_ARCH2)+
			(["R","A"][ABSOLUTE_DISPARITY]) +
			(["NS","S8"][SYM8_SUB])+
			"WLAM"+str(WLOSS_LAMBDA)+
			"SLAM"+str(SLOSS_LAMBDA)+
			"SCLP"+str(SLOSS_CLIP)+
			(['_nG','_G'][SPREAD_CONVERGENCE])+
			(['_nI','_I'][INTER_CONVERGENCE]) +
			(['_nHF',"_HF"][HOR_FLIP]) +
			('_CP'+str(DISP_DIFF_CAP)) +
			('_S'+str(DISP_DIFF_SLOPE))
			)
			##NN_LAYOUT1 = qcstereo_network.NN_LAYOUTS[NET_ARCH1]
			##NN_LAYOUT2 = qcstereo_network.NN_LAYOUTS[NET_ARCH2]
			# Tiff export slice labels
			SLICE_LABELS = ["nn_out_ext","hier_out_ext","gt_disparity","gt_strength"]#,
			# "cutcorn_cost_nw","cutcorn_cost",
			# "gt-avg_dist","avg8_disp","gt_disp","out-avg"]
			##############################################################################
			cluster_size = (2 * CLUSTER_RADIUS + 1) * (2 * CLUSTER_RADIUS + 1)
			center_tile_index = 2 * CLUSTER_RADIUS * (CLUSTER_RADIUS + 1)
			qsf.prepareFiles(dirs,
			files,
			suffix = SUFFIX)

			print ("Copying config files to results directory:\n ('%s' -> '%s')"%(conf_file,dirs['result']))
			try:
			os.makedirs(dirs['result'])
			except:
			pass

			shutil.copy2(conf_file,dirs['result'])
			LOGPATH = os.path.join(dirs['result'],LOGFILE)

			image_data = qsf.initImageData( # just use image_data[0]
			files = files,
			max_imgs = MAX_IMGS_IN_MEM,
			cluster_radius = 0, # CLUSTER_RADIUS,
			tile_layers = TILE_LAYERS,
			tile_side = TILE_SIDE,
			width = IMG_WIDTH,
			replace_nans = True,
			infer = True,
			keep_gt = True) # to generate same output files

			cluster_radius = CLUSTER_RADIUS

			ROOT_PATH = './attic/infer_qcds_graph'+SUFFIX+"/" # for tensorboard

			try:
			os.makedirs(os.path.dirname(files['inference']))
			print ("Created directory ",os.path.dirname(files['inference']))
			except:
			pass

			with tf.Session() as sess:
			infer_saver = tf.train.import_meta_graph(files["inference"]+'.meta')
			graph=tf.get_default_graph()
			ph_corr2d = graph.get_tensor_by_name('ph_corr2d:0')
			ph_target_disparity = graph.get_tensor_by_name('ph_target_disparity:0')
			ph_ntile = graph.get_tensor_by_name('ph_ntile:0')
			ph_ntile_out = graph.get_tensor_by_name('ph_ntile_out:0')

			stage1done = graph.get_tensor_by_name('Disparity_net/stage1done:0') #<tf.Operation 'Siam_net/stage1done' type=Const>,
			stage2_out_sparse = graph.get_tensor_by_name('Disparity_net/stage2_out_sparse:0')#not found
			if not USE_SPARSE_ONLY: #Does it reduce the graph size?
			stage2_out_full = graph.get_tensor_by_name('Disparity_net/stage2_out_full:0')

			sess.run(tf.global_variables_initializer())
			sess.run(tf.local_variables_initializer())

			infer_saver.restore(sess, files["inference"]) # after initializers, of course

			merged = tf.summary.merge_all()
			writer = tf.summary.FileWriter(ROOT_PATH, sess.graph)
			lf = None
			if LOGPATH:
			lf=open(LOGPATH,"w") #overwrite previous (or make it "a"?

			for nimg,_ in enumerate(image_data):
			dataset_img = qsf.readImageData(
			image_data = image_data,
			files = files,
			indx = nimg,
			cluster_radius = 0, # CLUSTER_RADIUS,
			tile_layers = TILE_LAYERS,
			tile_side = TILE_SIDE,
			width = IMG_WIDTH,
			replace_nans = True,
			infer = True,
			keep_gt = True) # to generate same output files
			img_corr2d = dataset_img['corr2d'] # [?,324)
			img_target = dataset_img['target_disparity'] # [?,324)
			img_ntile = dataset_img['ntile'].reshape([-1])
			#run first stage network
			qsf.print_time("Running inferred model, stage1", end=" ")
			_ = sess.run([stage1done],
			feed_dict={ph_corr2d: img_corr2d,
			ph_target_disparity: img_target,
			ph_ntile: img_ntile })
			qsf.print_time("Done.")
			qsf.print_time("Running inferred model, stage2", end=" ")
			disp_out, = sess.run([stage2_out_sparse],
			feed_dict={ph_ntile_out: img_ntile })
			qsf.print_time("Done.")
			result_file = files['result'][nimg].replace('.npy','-infer.npy') #not to overwrite training result files that are more complete
			try:
			os.makedirs(os.path.dirname(result_file))
			except:
			pass
			rslt = np.concatenate(
			[disp_out.reshape(-1,1),
			dataset_img['t_disps'], #t_disps[ntest],
			dataset_img['gtruths'], # gtruths[ntest],
			],1)
			np.save(result_file, rslt.reshape(HEIGHT,WIDTH,-1))
			rslt = qsf.eval_results(result_file, ABSOLUTE_DISPARITY, radius=CLUSTER_RADIUS, logfile=lf) # (re-loads results). Only uses first 4 layers
			if SAVE_TIFFS:
			qsf.result_npy_to_tiff(result_file, ABSOLUTE_DISPARITY, fix_nan = True,labels=SLICE_LABELS, logfile=lf)
			"""
			Remove dataset_img (if it is not [0] to reduce memory footprint
			"""
			image_data[nimg] = None

			if lf:
			lf.close()
			writer.close()

qcstereo_conf_sample.xml

+7 −0

Original line number	Original line	Diff line number	Diff line
	@@ -131,6 +131,10 @@
	<checkpoints>		<checkpoints>
	"tf_data_5x5_main_13_heur/checkpoints"		"tf_data_5x5_main_13_heur/checkpoints"
	</checkpoints>		</checkpoints>
			<inference>
			"tf_data_5x5_main_13_heur/inference"
			</inference>

	</directories>		</directories>
	<files>		<files>
	<train_lvar>		<train_lvar>
	@@ -330,6 +334,9 @@
	<checkpoints>		<checkpoints>
	"model_checkpoints"		"model_checkpoints"
	</checkpoints>		</checkpoints>
			<inference>
			"model"
			</inference>

	</files>		</files>
	</properties>		</properties>

qcstereo_functions.py

+8 −0

Original line number	Original line	Diff line number	Diff line
	@@ -118,11 +118,19 @@ def prepareFiles(dirs, files, suffix):
	for i, path in enumerate(files['images']):		for i, path in enumerate(files['images']):
	result_files.append(os.path.join(dirs['result'], path+"_"+suffix+'.npy'))		result_files.append(os.path.join(dirs['result'], path+"_"+suffix+'.npy'))
	files['result'] = result_files		files['result'] = result_files

	if not 'checkpoints' in files:		if not 'checkpoints' in files:
	files['checkpoints'] = 'checkpoints'		files['checkpoints'] = 'checkpoints'
	if not 'checkpoints' in dirs:		if not 'checkpoints' in dirs:
	dirs['checkpoints'] = dirs['result']		dirs['checkpoints'] = dirs['result']
	files['checkpoints'] = os.path.join(dirs['checkpoints'], files['checkpoints'])		files['checkpoints'] = os.path.join(dirs['checkpoints'], files['checkpoints'])

			if not 'inference' in files:
			files['inference'] = 'inference'
			if not 'inference' in dirs:
			dirs['inference'] = dirs['result']
			files['inference'] = os.path.join(dirs['inference'], files['inference'])

	if not 'figures' in dirs:		if not 'figures' in dirs:
	dirs['figures'] = os.path.join(dirs['result'],"figs")		dirs['figures'] = os.path.join(dirs['result'],"figs")

Admin message