preparing to feed nn

pack_tile.py

@@ -67,3 +67,41 @@ def pack_tile(tile,lut):
     out = np.append(out,layer)
 
   return out
+
+# tiles are already packed
+def get_tile_with_neighbors(tiles,i,j,radius):
+
+  out = np.array([])
+  # max
+  Y,X = tiles.shape[0:2]
+  #print(str(Y)+" "+str(X))
+
+  for k in range(2*radius+1):
+
+    y = i+k-radius
+    if   y<0:
+      y = 0
+    elif y>(Y-1):
+      y = Y-1
+
+    for l in range(2*radius+1):
+
+      x = j+l-radius
+      if   x<0:
+        x = 0
+      elif x>(X-1):
+        x = X-1
+
+      out = np.append(out,tiles[y,x])
+
+  return out
+
+
+
+
+
+
+
+
+
+

test_nn_feed.py

@@ -8,6 +8,8 @@ __email__     = "oleg@elphel.com"
 Open all tiffs in a folder, combine a single tiff from randomly selected
 tiles from originals
 '''
+import tensorflow as tf
+#import tensorflow.contrib.slim as slim
 
 from PIL import Image
 
@@ -21,32 +23,49 @@ import pack_tile as pile
 import numpy as np
 import itertools
 
+import time
+
+sys.exit()
+
+#http://stackoverflow.com/questions/287871/print-in-terminal-with-colors-using-python
+class bcolors:
+    HEADER = '\033[95m'
+    OKBLUE = '\033[94m'
+    OKGREEN = '\033[92m'
+    WARNING = '\033[38;5;214m'
+    FAIL = '\033[91m'
+    ENDC = '\033[0m'
+    BOLD = '\033[1m'
+    BOLDWHITE = '\033[1;37m'
+    UNDERLINE = '\033[4m'
+
+
+def print_time():
+  print(bcolors.BOLDWHITE+"time: "+str(time.time())+bcolors.ENDC)
+
 # USAGE: python3 test_3.py some-path
 
 VALUES_LAYER_NAME = 'other'
+LAYERS_OF_INTEREST = ['diagm-pair', 'diago-pair', 'hor-pairs', 'vert-pairs']
+RADIUS = 1
 
 try:
   src = sys.argv[1]
 except IndexError:
   src = "."
 
+print_time()
+
 tlist = glob.glob(src+"/*.tiff")
 
-print("Found "+str(len(tlist))+" tiff files:")
+print("Found "+str(len(tlist))+" preprocessed tiff files:")
 print("\n".join(tlist))
-
+print_time()
 ''' WARNING, assuming:
       - timestamps and part of names match
       - layer order and names are identical
 '''
 
-# CONSTANTS
-
-RADIUS = 1
-LAYERS_OF_INTEREST = ['diagm-pair','diago-pair']
-
-
-
 # open the first one to get dimensions and other info
 tiff = ijt.imagej_tiff(tlist[0])
 #del tlist[0]
@@ -58,15 +77,18 @@ tiff = ijt.imagej_tiff(tlist[0])
 labels = tiff.labels.copy()
 labels.remove(VALUES_LAYER_NAME)
 
-print("Image data layers: "+str(labels))
+print("Image data layers:  "+str(labels))
+print("Layers of interest: "+str(LAYERS_OF_INTEREST))
 print("Values layer: "+str([VALUES_LAYER_NAME]))
 
 # create copies
 tiles  = np.copy(tiff.getstack(labels,shape_as_tiles=True))
-tiles_bkp = np.copy(tiles)
 values = np.copy(tiff.getvalues(label=VALUES_LAYER_NAME))
 
-print("Tiled tiff shape: "+str(tiles.shape))
+#gt = values[:,:,1:3]
+
+print("Mixed tiled input data shape: "+str(tiles.shape))
+#print_time()
 
 # now generate a layer of indices to get other tiles
 indices = np.random.random_integers(0,len(tlist)-1,size=(tiles.shape[0],tiles.shape[1]))
@@ -97,7 +119,8 @@ for i in range(1,len(tlist)):
 for i in range(1,len(shuffle_counter)):
   shuffle_counter[0] -= shuffle_counter[i]
 
-print("Tiffs shuffle counter = "+str(shuffle_counter))
+print("Tiff files parts count in the mixed input = "+str(shuffle_counter))
+print_time()
 
 # test later
 
@@ -107,7 +130,77 @@ print("Tiffs shuffle counter = "+str(shuffle_counter))
 # Parse packing table
 # packing table name
 ptab_name = "tile_packing_table.xml"
-pt = pile.PackingTable(ptab_name,LAYERS_OF_INTEREST).lut
+ptab = pile.PackingTable(ptab_name,LAYERS_OF_INTEREST).lut
+
+# might not need it because going to loop through anyway
+packed_tiles = np.array([[pile.pack_tile(tiles[i,j],ptab) for j in range(tiles.shape[1])] for i in range(tiles.shape[0])])
+
+packed_tiles = np.dstack((packed_tiles,values[:,:,0]))
+
+print("Packed (81x4 -> 1x(25*4+1)) tiled input shape: "+str(packed_tiles.shape))
+print_time()
+
+#for i in range(tiles.shape[0]):
+#  for j in range(tiles.shape[1]):
+#    nn_input = pile.get_tile_with_neighbors(tiles,i,j,RADIUS)
+#    print("tile: "+str(i)+", "+str(j)+": shape = "+str(nn_input.shape))
+#print_time()
+
+result_dir = './result/'
+save_freq = 500
+
+def lrelu(x):
+    return tf.maximum(x*0.2,x)
+
+def network(input):
+
+  fc1 = slim.fully_connected(input,42,activation_fn=lrelu,scope='g_fc1')
+  fc2 = slim.fully_connected(fc1,  21,activation_fn=lrelu,scope='g_fc2')
+  fc3 = slim.fully_connected(fc2,   1,activation_fn=lrelu,scope='g_fc3')
+
+  return fc3
+
+
+sess = tf.session()
+
+in_tile = tf.placeholder(tf.float32,[None,None,101])
+gt      = tf.placeholder(tf.float32,[None,None,2])
+out = network(in_tile)
+
+G_loss = tf.reduce_mean(tf.abs(out[:,0]-gt[:,0]))
+t_vars=tf.trainable_variables()
+lr=tf.placeholder(tf.float32)
+G_opt=tf.train.AdamOptimizer(learning_rate=lr).minimize(G_loss,var_list=[var for var in t_vars if var.name.startswith('g_')])
+
+saver=tf.train.Saver()
+sess.run(tf.global_variables_initializer())
+ckpt=tf.train.get_checkpoint_state(checkpoint_dir)
+
+if ckpt:
+  print('loaded '+ckpt.model_checkpoint_path)
+  saver.restore(sess,ckpt.model_checkpoint_path)
+
+
+allfolders = glob.glob('./result/*0')
+lastepoch = 0
+for folder in allfolders:
+  lastepoch = np.maximum(lastepoch, int(folder[-4:]))
+
+#g_loss = np.zeros((,1))
+
+learning_rate = 1e-4
+for epoch in range(lastepoch,4001):
+  if os.path.isdir("result/%04d"%epoch):
+    continue
+  cnt=0
+  if epoch > 2000:
+    learning_rate = 1e-5
+
+  for ind in np.random.permutation(tiles.shape[0]*tiles.shape[1]):
+
+    input_patch = tiles[i,j]
+    gt_patch = values[i,j,1:2]
+
 
 
 

test_tf_setup.py

+#!/usr/bin/env python3
+
+__copyright__ = "Copyright 2018, Elphel, Inc."
+__license__   = "GPL-3.0+"
+__email__     = "oleg@elphel.com"
+
+'''
+Test:
+  nvidia graphic card
+  cuda installation
+  tensorflow
+
+Comment:
+  With nvidia + tensorflow - any software update casually breaks everything
+
+'''
+
+import subprocess
+import re
+
+#http://stackoverflow.com/questions/287871/print-in-terminal-with-colors-using-python
+class bcolors:
+    HEADER = '\033[95m'
+    OKBLUE = '\033[94m'
+    OKGREEN = '\033[92m'
+    WARNING = '\033[38;5;214m'
+    FAIL = '\033[91m'
+    ENDC = '\033[0m'
+    BOLD = '\033[1m'
+    BOLDWHITE = '\033[1;37m'
+    UNDERLINE = '\033[4m'
+
+
+
+# STEP 1: print nvidia model
+print(bcolors.BOLDWHITE+"NVIDIA devices:"+bcolors.ENDC)
+
+p = subprocess.run("lspci | grep NVIDIA",shell=True,stdout=subprocess.PIPE)
+out = p.stdout.strip().decode()
+if len(out)==0:
+  print(bcolors.FAIL+" not found (try 'lspci')"+bcolors.ENDC)
+else:
+  print(out)
+
+# STEP 2: nvidia driver version
+print(bcolors.BOLDWHITE+"NVIDIA driver version:"+bcolors.ENDC)
+p = subprocess.run("cat /proc/driver/nvidia/version",shell=True,stdout=subprocess.PIPE)
+out = p.stdout.strip().decode()
+print(out)
+
+# STEP 3: nvidia-smi - also some information about the graphics card and the driver
+print(bcolors.BOLDWHITE+"Some more info from 'nvidia-smi':"+bcolors.ENDC)
+p = subprocess.run("nvidia-smi",shell=True,stdout=subprocess.PIPE)
+out = p.stdout.strip().decode()
+print(out)
+
+
+print(bcolors.OKGREEN+"END"+bcolors.ENDC)
\ No newline at end of file

tile_packing_table.xml

@@ -53,4 +53,58 @@
     <tr row='23'>[(66,1.0), (67,1.0), (68,1.0)]</tr>
     <tr row='24'>[(74,1.0), (75,1.0), (76,1.0), (77,1.0), (78,1.0)]</tr>
   </table>
+  <table layer='hor-pairs'>
+    <tr row='0' >[(2,1.0),   (3,1.0),  (4,1.0),  (5,1.0), (6,1.0)]</tr>
+    <tr row='1' >[(10,1.0), (11,1.0), (19,1.0), (20,1.0)]</tr>
+    <tr row='2' >[(12,1.0), (13,1.0), (14,1.0)]</tr>
+    <tr row='3' >[(15,1.0), (16,1.0), (24,1.0), (25, 1.0)]</tr>
+    <tr row='4' >[(18,1.0), (27,1.0), (36,1.0), (45,1.0), (54,1.0)]</tr>
+    <tr row='5' >[(21,1.0), (22,1.0), (23,1.0)]</tr>
+    <tr row='6' >[(26,1.0), (35,1.0), (44,1.0), (53,1.0), (62,1.0)]</tr>
+    <tr row='7' >[(28,1.0), (37,1.0), (46,1.0)]</tr>
+    <tr row='8' >[(29,1.0), (38,1.0), (47,1.0)]</tr>
+    <tr row='9' >[(30,1.0)]</tr>
+    <tr row='10'>[(31,1.0)]</tr>
+    <tr row='11'>[(32,1.0)]</tr>
+    <tr row='12'>[(33,1.0), (42,1.0), (51,1.0)]</tr>
+    <tr row='13'>[(34,1.0), (43,1.0), (52,1.0)]</tr>
+    <tr row='14'>[(39,1.0)]</tr>
+    <tr row='15'>[(40,1.0)]</tr>
+    <tr row='16'>[(41,1.0)]</tr>
+    <tr row='17'>[(48,1.0)]</tr>
+    <tr row='18'>[(49,1.0)]</tr>
+    <tr row='19'>[(50,1.0)]</tr>
+    <tr row='20'>[(55,1.0), (56,1.0), (64,1.0), (65,1.0)]</tr>
+    <tr row='21'>[(57,1.0), (58,1.0), (59,1.0)]</tr>
+    <tr row='22'>[(60,1.0), (61,1.0), (69,1.0), (70,1.0)]</tr>
+    <tr row='23'>[(66,1.0), (67,1.0), (68,1.0)]</tr>
+    <tr row='24'>[(74,1.0), (75,1.0), (76,1.0), (77,1.0), (78,1.0)]</tr>
+  </table>
+  <table layer='vert-pairs'>
+    <tr row='0' >[(2,1.0),   (3,1.0),  (4,1.0),  (5,1.0), (6,1.0)]</tr>
+    <tr row='1' >[(10,1.0), (11,1.0), (19,1.0), (20,1.0)]</tr>
+    <tr row='2' >[(12,1.0), (13,1.0), (14,1.0)]</tr>
+    <tr row='3' >[(15,1.0), (16,1.0), (24,1.0), (25, 1.0)]</tr>
+    <tr row='4' >[(18,1.0), (27,1.0), (36,1.0), (45,1.0), (54,1.0)]</tr>
+    <tr row='5' >[(21,1.0), (22,1.0), (23,1.0)]</tr>
+    <tr row='6' >[(26,1.0), (35,1.0), (44,1.0), (53,1.0), (62,1.0)]</tr>
+    <tr row='7' >[(28,1.0), (37,1.0), (46,1.0)]</tr>
+    <tr row='8' >[(29,1.0), (38,1.0), (47,1.0)]</tr>
+    <tr row='9' >[(30,1.0)]</tr>
+    <tr row='10'>[(31,1.0)]</tr>
+    <tr row='11'>[(32,1.0)]</tr>
+    <tr row='12'>[(33,1.0), (42,1.0), (51,1.0)]</tr>
+    <tr row='13'>[(34,1.0), (43,1.0), (52,1.0)]</tr>
+    <tr row='14'>[(39,1.0)]</tr>
+    <tr row='15'>[(40,1.0)]</tr>
+    <tr row='16'>[(41,1.0)]</tr>
+    <tr row='17'>[(48,1.0)]</tr>
+    <tr row='18'>[(49,1.0)]</tr>
+    <tr row='19'>[(50,1.0)]</tr>
+    <tr row='20'>[(55,1.0), (56,1.0), (64,1.0), (65,1.0)]</tr>
+    <tr row='21'>[(57,1.0), (58,1.0), (59,1.0)]</tr>
+    <tr row='22'>[(60,1.0), (61,1.0), (69,1.0), (70,1.0)]</tr>
+    <tr row='23'>[(66,1.0), (67,1.0), (68,1.0)]</tr>
+    <tr row='24'>[(74,1.0), (75,1.0), (76,1.0), (77,1.0), (78,1.0)]</tr>
+  </table>
 </Document>
\ No newline at end of file