Update examples

examples/execute_jupyter_notebook_server.py

@@ -4,10 +4,14 @@ import subprocess
 from copy import deepcopy
 import socket
 from tempfile import mkstemp
+# make sure we have jupter in the auto requirements
 import jupyter
 from trains import Task
 
 
+# set default docker image, with network configuration
+os.environ['TRAINS_DOCKER_IMAGE'] = 'nvidia/cuda --network host'
+
 # initialize TRAINS
 task = Task.init(project_name='examples', task_name='Remote Jupyter NoteBook')
 
@@ -21,33 +25,45 @@ for key in os.environ:
     if key.startswith('TRAINS') and key not in preserve:
         env.pop(key, None)
 
+# Add jupyter server base folder
+param = {
+    'jupyter_server_base_directory': '',
+}
+task.connect(param)
+
 # execute jupyter notebook
 fd, local_filename = mkstemp()
-print('Running Jupyter Notebook Server on {} [{}]'.format(socket.gethostname(), socket.gethostbyname(socket.gethostname())))
+cwd = os.path.expandvars(os.path.expanduser(param['jupyter_server_base_directory'])) \
-process = subprocess.Popen([sys.executable, '-m', 'jupyter', 'notebook'], env=env, stdout=fd, stderr=fd)
+    if param['jupyter_server_base_directory'] else os.getcwd()
+print('Running Jupyter Notebook Server on {} [{}] at {}'.format(socket.gethostname(),
+                                                                socket.gethostbyname(socket.gethostname()), cwd))
+process = subprocess.Popen([sys.executable, '-m', 'jupyter', 'notebook', '--no-browser', '--allow-root'],
+                           env=env, stdout=fd, stderr=fd, cwd=cwd)
 
 # print stdout/stderr
 prev_line_count = 0
-while True:
+process_running = True
+while process_running:
+    process_running = False
     try:
         process.wait(timeout=2.0 if prev_line_count == 0 else 15.0)
     except subprocess.TimeoutExpired:
-        with open(local_filename, "rt") as f:
+        process_running = True
-            # read new lines
-            new_lines = f.readlines()
-            if not new_lines:
-                continue
-            output = ''.join(new_lines)
-            print(output)
-            # update task comment with jupyter notebook server links
-            if prev_line_count == 0:
-                task.comment += '\n' + ''.join(line for line in new_lines if 'http://' in line or 'https://' in line)
-            prev_line_count += len(new_lines)
 
-        os.lseek(fd, 0, 0)
+    with open(local_filename, "rt") as f:
-        os.ftruncate(fd, 0)
+        # read new lines
-        continue
+        new_lines = f.readlines()
-    break
+        if not new_lines:
+            continue
+        output = ''.join(new_lines)
+        print(output)
+        # update task comment with jupyter notebook server links
+        if prev_line_count == 0:
+            task.comment += '\n' + ''.join(line for line in new_lines if 'http://' in line or 'https://' in line)
+        prev_line_count += len(new_lines)
+
+    os.lseek(fd, 0, 0)
+    os.ftruncate(fd, 0)
 
 # cleanup
 os.close(fd)

examples/tensorflow_mnist_with_summaries.py

@@ -151,7 +151,7 @@ def train():
     def feed_dict(train):
         """Make a TensorFlow feed_dict: maps data onto Tensor placeholders."""
         if train or FLAGS.fake_data:
-            xs, ys = mnist.train.next_batch(100, fake_data=FLAGS.fake_data)
+            xs, ys = mnist.train.next_batch(FLAGS.batch_size, fake_data=FLAGS.fake_data)
             k = FLAGS.dropout
         else:
             xs, ys = mnist.test.images, mnist.test.labels
@@ -165,7 +165,7 @@ def train():
             test_writer.add_summary(summary, i)
             print('Accuracy at step %s: %s' % (i, acc))
         else:  # Record train set summaries, and train
-            if i % 100 == 99:  # Record execution stats
+            if i % FLAGS.batch_size == FLAGS.batch_size - 1:  # Record execution stats
                 run_metadata = tf.RunMetadata()
                 summary, _ = sess.run([merged, train_step],
                                       feed_dict=feed_dict(True),
@@ -213,5 +213,7 @@ if __name__ == '__main__':
                         help='Summaries log directory')
     parser.add_argument('--save_path', default=os.path.join(tempfile.gettempdir(), "model.ckpt"),
                         help='Save the trained model under this path')
+    parser.add_argument('--batch_size', default=100,
+                        help='Batch size for training')
     FLAGS, unparsed = parser.parse_known_args()
     tf.app.run(main=main, argv=[sys.argv[0]] + unparsed)