Python model.Graph() Examples
The following are 12
code examples of model.Graph().
You can vote up the ones you like or vote down the ones you don't like,
and go to the original project or source file by following the links above each example.
You may also want to check out all available functions/classes of the module
model
, or try the search function
.
.
Example #1
| Source File: neural_programmer.py From DOTA_models with Apache License 2.0 | 4 votes |
|
def master(train_data, dev_data, utility):
#creates TF graph and calls trainer or evaluator
batch_size = utility.FLAGS.batch_size
model_dir = utility.FLAGS.output_dir + "/model" + utility.FLAGS.job_id + "/"
#create all paramters of the model
param_class = parameters.Parameters(utility)
params, global_step, init = param_class.parameters(utility)
key = "test" if (FLAGS.evaluator_job) else "train"
graph = model.Graph(utility, batch_size, utility.FLAGS.max_passes, mode=key)
graph.create_graph(params, global_step)
prev_dev_error = 0.0
final_loss = 0.0
final_accuracy = 0.0
#start session
with tf.Session() as sess:
sess.run(init.name)
sess.run(graph.init_op.name)
to_save = params.copy()
saver = tf.train.Saver(to_save, max_to_keep=500)
if (FLAGS.evaluator_job):
while True:
selected_models = {}
file_list = tf.gfile.ListDirectory(model_dir)
for model_file in file_list:
if ("checkpoint" in model_file or "index" in model_file or
"meta" in model_file):
continue
if ("data" in model_file):
model_file = model_file.split(".")[0]
model_step = int(
model_file.split("_")[len(model_file.split("_")) - 1])
selected_models[model_step] = model_file
file_list = sorted(selected_models.items(), key=lambda x: x[0])
if (len(file_list) > 0):
file_list = file_list[0:len(file_list) - 1]
print "list of models: ", file_list
for model_file in file_list:
model_file = model_file[1]
print "restoring: ", model_file
saver.restore(sess, model_dir + "/" + model_file)
model_step = int(
model_file.split("_")[len(model_file.split("_")) - 1])
print "evaluating on dev ", model_file, model_step
evaluate(sess, dev_data, batch_size, graph, model_step)
else:
ckpt = tf.train.get_checkpoint_state(model_dir)
print "model dir: ", model_dir
if (not (tf.gfile.IsDirectory(utility.FLAGS.output_dir))):
print "create dir: ", utility.FLAGS.output_dir
tf.gfile.MkDir(utility.FLAGS.output_dir)
if (not (tf.gfile.IsDirectory(model_dir))):
print "create dir: ", model_dir
tf.gfile.MkDir(model_dir)
Train(graph, utility, batch_size, train_data, sess, model_dir,
saver)
Example #2
| Source File: neural_programmer.py From yolo_v2 with Apache License 2.0 | 4 votes |
|
def master(train_data, dev_data, utility):
#creates TF graph and calls trainer or evaluator
batch_size = utility.FLAGS.batch_size
model_dir = utility.FLAGS.output_dir + "/model" + utility.FLAGS.job_id + "/"
#create all paramters of the model
param_class = parameters.Parameters(utility)
params, global_step, init = param_class.parameters(utility)
key = "test" if (FLAGS.evaluator_job) else "train"
graph = model.Graph(utility, batch_size, utility.FLAGS.max_passes, mode=key)
graph.create_graph(params, global_step)
prev_dev_error = 0.0
final_loss = 0.0
final_accuracy = 0.0
#start session
with tf.Session() as sess:
sess.run(init.name)
sess.run(graph.init_op.name)
to_save = params.copy()
saver = tf.train.Saver(to_save, max_to_keep=500)
if (FLAGS.evaluator_job):
while True:
selected_models = {}
file_list = tf.gfile.ListDirectory(model_dir)
for model_file in file_list:
if ("checkpoint" in model_file or "index" in model_file or
"meta" in model_file):
continue
if ("data" in model_file):
model_file = model_file.split(".")[0]
model_step = int(
model_file.split("_")[len(model_file.split("_")) - 1])
selected_models[model_step] = model_file
file_list = sorted(selected_models.items(), key=lambda x: x[0])
if (len(file_list) > 0):
file_list = file_list[0:len(file_list) - 1]
print "list of models: ", file_list
for model_file in file_list:
model_file = model_file[1]
print "restoring: ", model_file
saver.restore(sess, model_dir + "/" + model_file)
model_step = int(
model_file.split("_")[len(model_file.split("_")) - 1])
print "evaluating on dev ", model_file, model_step
evaluate(sess, dev_data, batch_size, graph, model_step)
else:
ckpt = tf.train.get_checkpoint_state(model_dir)
print "model dir: ", model_dir
if (not (tf.gfile.IsDirectory(utility.FLAGS.output_dir))):
print "create dir: ", utility.FLAGS.output_dir
tf.gfile.MkDir(utility.FLAGS.output_dir)
if (not (tf.gfile.IsDirectory(model_dir))):
print "create dir: ", model_dir
tf.gfile.MkDir(model_dir)
Train(graph, utility, batch_size, train_data, sess, model_dir,
saver)
Example #3
| Source File: neural_programmer.py From Gun-Detector with Apache License 2.0 | 4 votes |
|
def master(train_data, dev_data, utility):
#creates TF graph and calls trainer or evaluator
batch_size = utility.FLAGS.batch_size
model_dir = utility.FLAGS.output_dir + "/model" + utility.FLAGS.job_id + "/"
#create all paramters of the model
param_class = parameters.Parameters(utility)
params, global_step, init = param_class.parameters(utility)
key = "test" if (FLAGS.evaluator_job) else "train"
graph = model.Graph(utility, batch_size, utility.FLAGS.max_passes, mode=key)
graph.create_graph(params, global_step)
prev_dev_error = 0.0
final_loss = 0.0
final_accuracy = 0.0
#start session
with tf.Session() as sess:
sess.run(init.name)
sess.run(graph.init_op.name)
to_save = params.copy()
saver = tf.train.Saver(to_save, max_to_keep=500)
if (FLAGS.evaluator_job):
while True:
selected_models = {}
file_list = tf.gfile.ListDirectory(model_dir)
for model_file in file_list:
if ("checkpoint" in model_file or "index" in model_file or
"meta" in model_file):
continue
if ("data" in model_file):
model_file = model_file.split(".")[0]
model_step = int(
model_file.split("_")[len(model_file.split("_")) - 1])
selected_models[model_step] = model_file
file_list = sorted(selected_models.items(), key=lambda x: x[0])
if (len(file_list) > 0):
file_list = file_list[0:len(file_list) - 1]
print("list of models: ", file_list)
for model_file in file_list:
model_file = model_file[1]
print("restoring: ", model_file)
saver.restore(sess, model_dir + "/" + model_file)
model_step = int(
model_file.split("_")[len(model_file.split("_")) - 1])
print("evaluating on dev ", model_file, model_step)
evaluate(sess, dev_data, batch_size, graph, model_step)
else:
ckpt = tf.train.get_checkpoint_state(model_dir)
print("model dir: ", model_dir)
if (not (tf.gfile.IsDirectory(utility.FLAGS.output_dir))):
print("create dir: ", utility.FLAGS.output_dir)
tf.gfile.MkDir(utility.FLAGS.output_dir)
if (not (tf.gfile.IsDirectory(model_dir))):
print("create dir: ", model_dir)
tf.gfile.MkDir(model_dir)
Train(graph, utility, batch_size, train_data, sess, model_dir,
saver)
Example #4
| Source File: neural_programmer.py From Action_Recognition_Zoo with MIT License | 4 votes |
|
def master(train_data, dev_data, utility):
#creates TF graph and calls trainer or evaluator
batch_size = utility.FLAGS.batch_size
model_dir = utility.FLAGS.output_dir + "/model" + utility.FLAGS.job_id + "/"
#create all paramters of the model
param_class = parameters.Parameters(utility)
params, global_step, init = param_class.parameters(utility)
key = "test" if (FLAGS.evaluator_job) else "train"
graph = model.Graph(utility, batch_size, utility.FLAGS.max_passes, mode=key)
graph.create_graph(params, global_step)
prev_dev_error = 0.0
final_loss = 0.0
final_accuracy = 0.0
#start session
with tf.Session() as sess:
sess.run(init.name)
sess.run(graph.init_op.name)
to_save = params.copy()
saver = tf.train.Saver(to_save, max_to_keep=500)
if (FLAGS.evaluator_job):
while True:
selected_models = {}
file_list = tf.gfile.ListDirectory(model_dir)
for model_file in file_list:
if ("checkpoint" in model_file or "index" in model_file or
"meta" in model_file):
continue
if ("data" in model_file):
model_file = model_file.split(".")[0]
model_step = int(
model_file.split("_")[len(model_file.split("_")) - 1])
selected_models[model_step] = model_file
file_list = sorted(selected_models.items(), key=lambda x: x[0])
if (len(file_list) > 0):
file_list = file_list[0:len(file_list) - 1]
print "list of models: ", file_list
for model_file in file_list:
model_file = model_file[1]
print "restoring: ", model_file
saver.restore(sess, model_dir + "/" + model_file)
model_step = int(
model_file.split("_")[len(model_file.split("_")) - 1])
print "evaluating on dev ", model_file, model_step
evaluate(sess, dev_data, batch_size, graph, model_step)
else:
ckpt = tf.train.get_checkpoint_state(model_dir)
print "model dir: ", model_dir
if (not (tf.gfile.IsDirectory(model_dir))):
print "create dir: ", model_dir
tf.gfile.MkDir(model_dir)
Train(graph, utility, batch_size, train_data, sess, model_dir,
saver)
Example #5
| Source File: neural_programmer.py From ECO-pytorch with BSD 2-Clause "Simplified" License | 4 votes |
|
def master(train_data, dev_data, utility):
#creates TF graph and calls trainer or evaluator
batch_size = utility.FLAGS.batch_size
model_dir = utility.FLAGS.output_dir + "/model" + utility.FLAGS.job_id + "/"
#create all paramters of the model
param_class = parameters.Parameters(utility)
params, global_step, init = param_class.parameters(utility)
key = "test" if (FLAGS.evaluator_job) else "train"
graph = model.Graph(utility, batch_size, utility.FLAGS.max_passes, mode=key)
graph.create_graph(params, global_step)
prev_dev_error = 0.0
final_loss = 0.0
final_accuracy = 0.0
#start session
with tf.Session() as sess:
sess.run(init.name)
sess.run(graph.init_op.name)
to_save = params.copy()
saver = tf.train.Saver(to_save, max_to_keep=500)
if (FLAGS.evaluator_job):
while True:
selected_models = {}
file_list = tf.gfile.ListDirectory(model_dir)
for model_file in file_list:
if ("checkpoint" in model_file or "index" in model_file or
"meta" in model_file):
continue
if ("data" in model_file):
model_file = model_file.split(".")[0]
model_step = int(
model_file.split("_")[len(model_file.split("_")) - 1])
selected_models[model_step] = model_file
file_list = sorted(selected_models.items(), key=lambda x: x[0])
if (len(file_list) > 0):
file_list = file_list[0:len(file_list) - 1]
print "list of models: ", file_list
for model_file in file_list:
model_file = model_file[1]
print "restoring: ", model_file
saver.restore(sess, model_dir + "/" + model_file)
model_step = int(
model_file.split("_")[len(model_file.split("_")) - 1])
print "evaluating on dev ", model_file, model_step
evaluate(sess, dev_data, batch_size, graph, model_step)
else:
ckpt = tf.train.get_checkpoint_state(model_dir)
print "model dir: ", model_dir
if (not (tf.gfile.IsDirectory(model_dir))):
print "create dir: ", model_dir
tf.gfile.MkDir(model_dir)
Train(graph, utility, batch_size, train_data, sess, model_dir,
saver)
Example #6
| Source File: neural_programmer.py From hands-detection with MIT License | 4 votes |
|
def master(train_data, dev_data, utility):
#creates TF graph and calls trainer or evaluator
batch_size = utility.FLAGS.batch_size
model_dir = utility.FLAGS.output_dir + "/model" + utility.FLAGS.job_id + "/"
#create all paramters of the model
param_class = parameters.Parameters(utility)
params, global_step, init = param_class.parameters(utility)
key = "test" if (FLAGS.evaluator_job) else "train"
graph = model.Graph(utility, batch_size, utility.FLAGS.max_passes, mode=key)
graph.create_graph(params, global_step)
prev_dev_error = 0.0
final_loss = 0.0
final_accuracy = 0.0
#start session
with tf.Session() as sess:
sess.run(init.name)
sess.run(graph.init_op.name)
to_save = params.copy()
saver = tf.train.Saver(to_save, max_to_keep=500)
if (FLAGS.evaluator_job):
while True:
selected_models = {}
file_list = tf.gfile.ListDirectory(model_dir)
for model_file in file_list:
if ("checkpoint" in model_file or "index" in model_file or
"meta" in model_file):
continue
if ("data" in model_file):
model_file = model_file.split(".")[0]
model_step = int(
model_file.split("_")[len(model_file.split("_")) - 1])
selected_models[model_step] = model_file
file_list = sorted(selected_models.items(), key=lambda x: x[0])
if (len(file_list) > 0):
file_list = file_list[0:len(file_list) - 1]
print "list of models: ", file_list
for model_file in file_list:
model_file = model_file[1]
print "restoring: ", model_file
saver.restore(sess, model_dir + "/" + model_file)
model_step = int(
model_file.split("_")[len(model_file.split("_")) - 1])
print "evaluating on dev ", model_file, model_step
evaluate(sess, dev_data, batch_size, graph, model_step)
else:
ckpt = tf.train.get_checkpoint_state(model_dir)
print "model dir: ", model_dir
if (not (tf.gfile.IsDirectory(utility.FLAGS.output_dir))):
print "create dir: ", utility.FLAGS.output_dir
tf.gfile.MkDir(utility.FLAGS.output_dir)
if (not (tf.gfile.IsDirectory(model_dir))):
print "create dir: ", model_dir
tf.gfile.MkDir(model_dir)
Train(graph, utility, batch_size, train_data, sess, model_dir,
saver)
Example #7
| Source File: neural_programmer.py From object_detection_kitti with Apache License 2.0 | 4 votes |
|
def master(train_data, dev_data, utility):
#creates TF graph and calls trainer or evaluator
batch_size = utility.FLAGS.batch_size
model_dir = utility.FLAGS.output_dir + "/model" + utility.FLAGS.job_id + "/"
#create all paramters of the model
param_class = parameters.Parameters(utility)
params, global_step, init = param_class.parameters(utility)
key = "test" if (FLAGS.evaluator_job) else "train"
graph = model.Graph(utility, batch_size, utility.FLAGS.max_passes, mode=key)
graph.create_graph(params, global_step)
prev_dev_error = 0.0
final_loss = 0.0
final_accuracy = 0.0
#start session
with tf.Session() as sess:
sess.run(init.name)
sess.run(graph.init_op.name)
to_save = params.copy()
saver = tf.train.Saver(to_save, max_to_keep=500)
if (FLAGS.evaluator_job):
while True:
selected_models = {}
file_list = tf.gfile.ListDirectory(model_dir)
for model_file in file_list:
if ("checkpoint" in model_file or "index" in model_file or
"meta" in model_file):
continue
if ("data" in model_file):
model_file = model_file.split(".")[0]
model_step = int(
model_file.split("_")[len(model_file.split("_")) - 1])
selected_models[model_step] = model_file
file_list = sorted(selected_models.items(), key=lambda x: x[0])
if (len(file_list) > 0):
file_list = file_list[0:len(file_list) - 1]
print "list of models: ", file_list
for model_file in file_list:
model_file = model_file[1]
print "restoring: ", model_file
saver.restore(sess, model_dir + "/" + model_file)
model_step = int(
model_file.split("_")[len(model_file.split("_")) - 1])
print "evaluating on dev ", model_file, model_step
evaluate(sess, dev_data, batch_size, graph, model_step)
else:
ckpt = tf.train.get_checkpoint_state(model_dir)
print "model dir: ", model_dir
if (not (tf.gfile.IsDirectory(utility.FLAGS.output_dir))):
print "create dir: ", utility.FLAGS.output_dir
tf.gfile.MkDir(utility.FLAGS.output_dir)
if (not (tf.gfile.IsDirectory(model_dir))):
print "create dir: ", model_dir
tf.gfile.MkDir(model_dir)
Train(graph, utility, batch_size, train_data, sess, model_dir,
saver)
Example #8
| Source File: neural_programmer.py From object_detection_with_tensorflow with MIT License | 4 votes |
|
def master(train_data, dev_data, utility):
#creates TF graph and calls trainer or evaluator
batch_size = utility.FLAGS.batch_size
model_dir = utility.FLAGS.output_dir + "/model" + utility.FLAGS.job_id + "/"
#create all paramters of the model
param_class = parameters.Parameters(utility)
params, global_step, init = param_class.parameters(utility)
key = "test" if (FLAGS.evaluator_job) else "train"
graph = model.Graph(utility, batch_size, utility.FLAGS.max_passes, mode=key)
graph.create_graph(params, global_step)
prev_dev_error = 0.0
final_loss = 0.0
final_accuracy = 0.0
#start session
with tf.Session() as sess:
sess.run(init.name)
sess.run(graph.init_op.name)
to_save = params.copy()
saver = tf.train.Saver(to_save, max_to_keep=500)
if (FLAGS.evaluator_job):
while True:
selected_models = {}
file_list = tf.gfile.ListDirectory(model_dir)
for model_file in file_list:
if ("checkpoint" in model_file or "index" in model_file or
"meta" in model_file):
continue
if ("data" in model_file):
model_file = model_file.split(".")[0]
model_step = int(
model_file.split("_")[len(model_file.split("_")) - 1])
selected_models[model_step] = model_file
file_list = sorted(selected_models.items(), key=lambda x: x[0])
if (len(file_list) > 0):
file_list = file_list[0:len(file_list) - 1]
print "list of models: ", file_list
for model_file in file_list:
model_file = model_file[1]
print "restoring: ", model_file
saver.restore(sess, model_dir + "/" + model_file)
model_step = int(
model_file.split("_")[len(model_file.split("_")) - 1])
print "evaluating on dev ", model_file, model_step
evaluate(sess, dev_data, batch_size, graph, model_step)
else:
ckpt = tf.train.get_checkpoint_state(model_dir)
print "model dir: ", model_dir
if (not (tf.gfile.IsDirectory(utility.FLAGS.output_dir))):
print "create dir: ", utility.FLAGS.output_dir
tf.gfile.MkDir(utility.FLAGS.output_dir)
if (not (tf.gfile.IsDirectory(model_dir))):
print "create dir: ", model_dir
tf.gfile.MkDir(model_dir)
Train(graph, utility, batch_size, train_data, sess, model_dir,
saver)
Example #9
| Source File: neural_programmer.py From HumanRecognition with MIT License | 4 votes |
|
def master(train_data, dev_data, utility):
#creates TF graph and calls trainer or evaluator
batch_size = utility.FLAGS.batch_size
model_dir = utility.FLAGS.output_dir + "/model" + utility.FLAGS.job_id + "/"
#create all paramters of the model
param_class = parameters.Parameters(utility)
params, global_step, init = param_class.parameters(utility)
key = "test" if (FLAGS.evaluator_job) else "train"
graph = model.Graph(utility, batch_size, utility.FLAGS.max_passes, mode=key)
graph.create_graph(params, global_step)
prev_dev_error = 0.0
final_loss = 0.0
final_accuracy = 0.0
#start session
with tf.Session() as sess:
sess.run(init.name)
sess.run(graph.init_op.name)
to_save = params.copy()
saver = tf.train.Saver(to_save, max_to_keep=500)
if (FLAGS.evaluator_job):
while True:
selected_models = {}
file_list = tf.gfile.ListDirectory(model_dir)
for model_file in file_list:
if ("checkpoint" in model_file or "index" in model_file or
"meta" in model_file):
continue
if ("data" in model_file):
model_file = model_file.split(".")[0]
model_step = int(
model_file.split("_")[len(model_file.split("_")) - 1])
selected_models[model_step] = model_file
file_list = sorted(selected_models.items(), key=lambda x: x[0])
if (len(file_list) > 0):
file_list = file_list[0:len(file_list) - 1]
print "list of models: ", file_list
for model_file in file_list:
model_file = model_file[1]
print "restoring: ", model_file
saver.restore(sess, model_dir + "/" + model_file)
model_step = int(
model_file.split("_")[len(model_file.split("_")) - 1])
print "evaluating on dev ", model_file, model_step
evaluate(sess, dev_data, batch_size, graph, model_step)
else:
ckpt = tf.train.get_checkpoint_state(model_dir)
print "model dir: ", model_dir
if (not (tf.gfile.IsDirectory(utility.FLAGS.output_dir))):
print "create dir: ", utility.FLAGS.output_dir
tf.gfile.MkDir(utility.FLAGS.output_dir)
if (not (tf.gfile.IsDirectory(model_dir))):
print "create dir: ", model_dir
tf.gfile.MkDir(model_dir)
Train(graph, utility, batch_size, train_data, sess, model_dir,
saver)
Example #10
| Source File: neural_programmer.py From g-tensorflow-models with Apache License 2.0 | 4 votes |
|
def master(train_data, dev_data, utility):
#creates TF graph and calls trainer or evaluator
batch_size = utility.FLAGS.batch_size
model_dir = utility.FLAGS.output_dir + "/model" + utility.FLAGS.job_id + "/"
#create all paramters of the model
param_class = parameters.Parameters(utility)
params, global_step, init = param_class.parameters(utility)
key = "test" if (FLAGS.evaluator_job) else "train"
graph = model.Graph(utility, batch_size, utility.FLAGS.max_passes, mode=key)
graph.create_graph(params, global_step)
prev_dev_error = 0.0
final_loss = 0.0
final_accuracy = 0.0
#start session
with tf.Session() as sess:
sess.run(init.name)
sess.run(graph.init_op.name)
to_save = params.copy()
saver = tf.train.Saver(to_save, max_to_keep=500)
if (FLAGS.evaluator_job):
while True:
selected_models = {}
file_list = tf.gfile.ListDirectory(model_dir)
for model_file in file_list:
if ("checkpoint" in model_file or "index" in model_file or
"meta" in model_file):
continue
if ("data" in model_file):
model_file = model_file.split(".")[0]
model_step = int(
model_file.split("_")[len(model_file.split("_")) - 1])
selected_models[model_step] = model_file
file_list = sorted(selected_models.items(), key=lambda x: x[0])
if (len(file_list) > 0):
file_list = file_list[0:len(file_list) - 1]
print("list of models: ", file_list)
for model_file in file_list:
model_file = model_file[1]
print("restoring: ", model_file)
saver.restore(sess, model_dir + "/" + model_file)
model_step = int(
model_file.split("_")[len(model_file.split("_")) - 1])
print("evaluating on dev ", model_file, model_step)
evaluate(sess, dev_data, batch_size, graph, model_step)
else:
ckpt = tf.train.get_checkpoint_state(model_dir)
print("model dir: ", model_dir)
if (not (tf.gfile.IsDirectory(utility.FLAGS.output_dir))):
print("create dir: ", utility.FLAGS.output_dir)
tf.gfile.MkDir(utility.FLAGS.output_dir)
if (not (tf.gfile.IsDirectory(model_dir))):
print("create dir: ", model_dir)
tf.gfile.MkDir(model_dir)
Train(graph, utility, batch_size, train_data, sess, model_dir,
saver)
Example #11
| Source File: neural_programmer.py From models with Apache License 2.0 | 4 votes |
|
def master(train_data, dev_data, utility):
#creates TF graph and calls trainer or evaluator
batch_size = utility.FLAGS.batch_size
model_dir = utility.FLAGS.output_dir + "/model" + utility.FLAGS.job_id + "/"
#create all paramters of the model
param_class = parameters.Parameters(utility)
params, global_step, init = param_class.parameters(utility)
key = "test" if (FLAGS.evaluator_job) else "train"
graph = model.Graph(utility, batch_size, utility.FLAGS.max_passes, mode=key)
graph.create_graph(params, global_step)
prev_dev_error = 0.0
final_loss = 0.0
final_accuracy = 0.0
#start session
with tf.Session() as sess:
sess.run(init.name)
sess.run(graph.init_op.name)
to_save = params.copy()
saver = tf.train.Saver(to_save, max_to_keep=500)
if (FLAGS.evaluator_job):
while True:
selected_models = {}
file_list = tf.gfile.ListDirectory(model_dir)
for model_file in file_list:
if ("checkpoint" in model_file or "index" in model_file or
"meta" in model_file):
continue
if ("data" in model_file):
model_file = model_file.split(".")[0]
model_step = int(
model_file.split("_")[len(model_file.split("_")) - 1])
selected_models[model_step] = model_file
file_list = sorted(selected_models.items(), key=lambda x: x[0])
if (len(file_list) > 0):
file_list = file_list[0:len(file_list) - 1]
print("list of models: ", file_list)
for model_file in file_list:
model_file = model_file[1]
print("restoring: ", model_file)
saver.restore(sess, model_dir + "/" + model_file)
model_step = int(
model_file.split("_")[len(model_file.split("_")) - 1])
print("evaluating on dev ", model_file, model_step)
evaluate(sess, dev_data, batch_size, graph, model_step)
else:
ckpt = tf.train.get_checkpoint_state(model_dir)
print("model dir: ", model_dir)
if (not (tf.gfile.IsDirectory(utility.FLAGS.output_dir))):
print("create dir: ", utility.FLAGS.output_dir)
tf.gfile.MkDir(utility.FLAGS.output_dir)
if (not (tf.gfile.IsDirectory(model_dir))):
print("create dir: ", model_dir)
tf.gfile.MkDir(model_dir)
Train(graph, utility, batch_size, train_data, sess, model_dir,
saver)
Example #12
| Source File: neural_programmer.py From multilabel-image-classification-tensorflow with MIT License | 4 votes |
|
def master(train_data, dev_data, utility):
#creates TF graph and calls trainer or evaluator
batch_size = utility.FLAGS.batch_size
model_dir = utility.FLAGS.output_dir + "/model" + utility.FLAGS.job_id + "/"
#create all paramters of the model
param_class = parameters.Parameters(utility)
params, global_step, init = param_class.parameters(utility)
key = "test" if (FLAGS.evaluator_job) else "train"
graph = model.Graph(utility, batch_size, utility.FLAGS.max_passes, mode=key)
graph.create_graph(params, global_step)
prev_dev_error = 0.0
final_loss = 0.0
final_accuracy = 0.0
#start session
with tf.Session() as sess:
sess.run(init.name)
sess.run(graph.init_op.name)
to_save = params.copy()
saver = tf.train.Saver(to_save, max_to_keep=500)
if (FLAGS.evaluator_job):
while True:
selected_models = {}
file_list = tf.gfile.ListDirectory(model_dir)
for model_file in file_list:
if ("checkpoint" in model_file or "index" in model_file or
"meta" in model_file):
continue
if ("data" in model_file):
model_file = model_file.split(".")[0]
model_step = int(
model_file.split("_")[len(model_file.split("_")) - 1])
selected_models[model_step] = model_file
file_list = sorted(selected_models.items(), key=lambda x: x[0])
if (len(file_list) > 0):
file_list = file_list[0:len(file_list) - 1]
print("list of models: ", file_list)
for model_file in file_list:
model_file = model_file[1]
print("restoring: ", model_file)
saver.restore(sess, model_dir + "/" + model_file)
model_step = int(
model_file.split("_")[len(model_file.split("_")) - 1])
print("evaluating on dev ", model_file, model_step)
evaluate(sess, dev_data, batch_size, graph, model_step)
else:
ckpt = tf.train.get_checkpoint_state(model_dir)
print("model dir: ", model_dir)
if (not (tf.gfile.IsDirectory(utility.FLAGS.output_dir))):
print("create dir: ", utility.FLAGS.output_dir)
tf.gfile.MkDir(utility.FLAGS.output_dir)
if (not (tf.gfile.IsDirectory(model_dir))):
print("create dir: ", model_dir)
tf.gfile.MkDir(model_dir)
Train(graph, utility, batch_size, train_data, sess, model_dir,
saver)
