def train(args):
datasets = range(4)
# Remove the leaveDataset from datasets
datasets.remove(args.leaveDataset)
# Create the SocialDataLoader object
data_loader = SocialDataLoader(args.batch_size,
args.seq_length,
args.maxNumPeds,
datasets,
forcePreProcess=True)
with open(os.path.join('save', 'social_config.pkl'), 'wb') as f:
pickle.dump(args, f)
# Create a SocialModel object with the arguments
model = SocialModel(args)
# Initialize a TensorFlow session
with tf.Session() as sess:
# Initialize all variables in the graph
sess.run(tf.initialize_all_variables())
# Initialize a saver that saves all the variables in the graph
saver = tf.train.Saver(tf.all_variables())
# summary_writer = tf.train.SummaryWriter('/tmp/lstm/logs', graph_def=sess.graph_def)
# For each epoch
for e in range(args.num_epochs):
# Assign the learning rate value for this epoch
sess.run(
tf.assign(model.lr, args.learning_rate * (args.decay_rate**e)))
# Reset the data pointers in the data_loader
data_loader.reset_batch_pointer()
# For each batch
for b in range(data_loader.num_batches):
# Tic
start = time.time()
# Get the source, target and dataset data for the next batch
# x, y are input and target data which are lists containing numpy arrays of size seq_length x maxNumPeds x 3
# d is the list of dataset indices from which each batch is generated (used to differentiate between datasets)
x, y, d = data_loader.next_batch()
# variable to store the loss for this batch
loss_batch = 0
# For each sequence in the batch
for batch in range(data_loader.batch_size):
# x_batch, y_batch and d_batch contains the source, target and dataset index data for
# seq_length long consecutive frames in the dataset
# x_batch, y_batch would be numpy arrays of size seq_length x maxNumPeds x 3
# d_batch would be a scalar identifying the dataset from which this sequence is extracted
x_batch, y_batch, d_batch = x[batch], y[batch], d[batch]
if d_batch == 0 and datasets[0] == 0:
dataset_data = [640, 480]
else:
dataset_data = [720, 576]
grid_batch = getSequenceGridMask(x_batch, dataset_data,
args.neighborhood_size,
args.grid_size)
# Feed the source, target data
feed = {
model.input_data: x_batch,
model.target_data: y_batch,
model.grid_data: grid_batch
}
train_loss, _ = sess.run([model.cost, model.train_op],
feed)
loss_batch += train_loss
end = time.time()
loss_batch = loss_batch / data_loader.batch_size
print(
"{}/{} (epoch {}), train_loss = {:.3f}, time/batch = {:.3f}"
.format(e * data_loader.num_batches + b,
args.num_epochs * data_loader.num_batches, e,
loss_batch, end - start))
# Save the model if the current epoch and batch number match the frequency
if (e * data_loader.num_batches +
b) % args.save_every == 0 and (
(e * data_loader.num_batches + b) > 0):
checkpoint_path = os.path.join('save', 'social_model.ckpt')
saver.save(sess,
checkpoint_path,
global_step=e * data_loader.num_batches + b)
print("model saved to {}".format(checkpoint_path))
def main():
parser = argparse.ArgumentParser()
# Observed length of the trajectory parameter
parser.add_argument('--obs_length',
type=int,
default=8,
help='Observed length of the trajectory')
# Predicted length of the trajectory parameter
parser.add_argument('--pred_length',
type=int,
default=12,
help='Predicted length of the trajectory')
# Test dataset
parser.add_argument('--test_dataset',
type=int,
default=0,
help='Dataset to be tested on')
# Parse the parameters
sample_args = parser.parse_args()
# Define the path for the config file for saved args
with open(os.path.join('save', 'social_config.pkl'), 'rb') as f:
saved_args = pickle.load(f)
# Create a SocialModel object with the saved_args and infer set to true
model = SocialModel(saved_args, True)
# Initialize a TensorFlow session
sess = tf.InteractiveSession()
# Initialize a saver
saver = tf.train.Saver()
# Get the checkpoint state for the model
ckpt = tf.train.get_checkpoint_state('save')
print('loading model: ', ckpt.model_checkpoint_path)
# Restore the model at the checkpoint
saver.restore(sess, ckpt.model_checkpoint_path)
# saver.restore(sess, 'save/social_model.ckpt-800')
# Dataset to get data from
dataset = [sample_args.test_dataset]
# Create a SocialDataLoader object with batch_size 1 and seq_length equal to observed_length + pred_length
data_loader = SocialDataLoader(
1, sample_args.pred_length + sample_args.obs_length,
saved_args.maxNumPeds, dataset, True)
# Reset all pointers of the data_loader
data_loader.reset_batch_pointer()
# Variable to maintain total error
total_error = 0
# For each batch
for b in range(data_loader.num_batches):
# Get the source, target and dataset data for the next batch
x, y, d = data_loader.next_batch()
# Batch size is 1
x_batch, y_batch, d_batch = x[0], y[0], d[0]
if d_batch == 0 and dataset[0] == 0:
dimensions = [640, 480]
else:
dimensions = [720, 576]
grid_batch = getSequenceGridMask(x_batch, dimensions,
saved_args.neighborhood_size,
saved_args.grid_size)
obs_traj = x_batch[:sample_args.obs_length]
obs_grid = grid_batch[:sample_args.obs_length]
# obs_traj is an array of shape obs_length x maxNumPeds x 3
complete_traj = model.sample(sess, obs_traj, obs_grid, dimensions,
x_batch, sample_args.pred_length)
# ipdb.set_trace()
# complete_traj is an array of shape (obs_length+pred_length) x maxNumPeds x 3
total_error += get_mean_error(complete_traj, x[0],
sample_args.obs_length,
saved_args.maxNumPeds)
print "Processed trajectory number : ", b, "out of ", data_loader.num_batches, " trajectories"
# Print the mean error across all the batches
print "Total mean error of the model is ", total_error / data_loader.num_batches
def train(args):
datasets = range(5)
# Remove the leaveDataset from datasets
datasets.remove(args.leaveDataset)
# Create the SocialDataLoader object
data_loader = SocialDataLoader(args.batch_size,
args.seq_length,
args.maxNumPeds,
datasets,
forcePreProcess=True,
infer=False)
# Log directory
log_directory = 'log/'
log_directory += str(args.leaveDataset) + '/'
# Logging files
log_file_curve = open(os.path.join(log_directory, 'log_curve.txt'), 'w')
log_file = open(os.path.join(log_directory, 'val.txt'), 'w')
# Save directory
save_directory = 'save/'
save_directory += str(args.leaveDataset) + '/'
with open(os.path.join(save_directory, 'social_config.pkl'), 'wb') as f:
pickle.dump(args, f)
# Create a SocialModel object with the arguments
model = SocialModel(args)
config = tf.ConfigProto()
config.gpu_options.allow_growth = True
config = tf.ConfigProto(
log_device_placement=True
) # Showing which device is allocated (in case of multiple GPUs)
config.gpu_options.per_process_gpu_memory_fraction = 0.8 # Allocating 20% of memory in each GPU with 0.5
# Initialize a TensorFlow session
with tf.Session() as sess:
# Initialize all variables in the graph
sess.run(tf.global_variables_initializer())
# Initialize a saver that saves all the variables in the graph
saver = tf.train.Saver(tf.global_variables(), max_to_keep=None)
# summary_writer = tf.train.SummaryWriter('/tmp/lstm/logs', graph_def=sess.graph_def)
print('Training begin')
best_val_loss = 100
best_epoch = 0
# For each epoch
for e in range(args.num_epochs):
# Assign the learning rate value for this epoch
sess.run(
tf.assign(model.lr, args.learning_rate * (args.decay_rate**e)))
# Reset the data pointers in the data_loader
data_loader.reset_batch_pointer(valid=False)
loss_epoch = 0
# For each batch
for b in range(data_loader.num_batches):
# Tic
start = time.time()
# Get the source, target and dataset data for the next batch
# x, y are input and target data which are lists containing numpy arrays of size seq_length x maxNumPeds x 3
# d is the list of dataset indices from which each batch is generated (used to differentiate between datasets)
x, y, d = data_loader.next_batch()
# variable to store the loss for this batch
loss_batch = 0
# For each sequence in the batch
for batch in range(data_loader.batch_size):
# x_batch, y_batch and d_batch contains the source, target and dataset index data for
# seq_length long consecutive frames in the dataset
# x_batch, y_batch would be numpy arrays of size seq_length x maxNumPeds x 3
# d_batch would be a scalar identifying the dataset from which this sequence is extracted
x_batch, y_batch, d_batch = x[batch], y[batch], d[batch]
if d_batch == 0 and datasets[0] == 0:
dataset_data = [640, 480]
else:
dataset_data = [720, 576]
grid_batch = getSequenceGridMask(x_batch, dataset_data,
args.neighborhood_size,
args.grid_size)
# Feed the source, target data
feed = {
model.input_data: x_batch,
model.target_data: y_batch,
model.grid_data: grid_batch
}
train_loss, _ = sess.run([model.cost, model.train_op],
feed)
loss_batch += train_loss
end = time.time()
loss_batch = loss_batch / data_loader.batch_size
loss_epoch += loss_batch
print(
"{}/{} (epoch {}), train_loss = {:.3f}, time/batch = {:.3f}"
.format(e * data_loader.num_batches + b,
args.num_epochs * data_loader.num_batches, e,
loss_batch, end - start))
# Save the model if the current epoch and batch number match the frequency
'''
if (e * data_loader.num_batches + b) % args.save_every == 0 and ((e * data_loader.num_batches + b) > 0):
checkpoint_path = os.path.join('save', 'social_model.ckpt')
saver.save(sess, checkpoint_path, global_step=e * data_loader.num_batches + b)
print("model saved to {}".format(checkpoint_path))
'''
loss_epoch /= data_loader.num_batches
log_file_curve.write(str(e) + ',' + str(loss_epoch) + ',')
print('*****************')
# Validation
data_loader.reset_batch_pointer(valid=True)
loss_epoch = 0
for b in range(data_loader.num_batches):
# Get the source, target and dataset data for the next batch
# x, y are input and target data which are lists containing numpy arrays of size seq_length x maxNumPeds x 3
# d is the list of dataset indices from which each batch is generated (used to differentiate between datasets)
x, y, d = data_loader.next_valid_batch()
# variable to store the loss for this batch
loss_batch = 0
# For each sequence in the batch
for batch in range(data_loader.batch_size):
# x_batch, y_batch and d_batch contains the source, target and dataset index data for
# seq_length long consecutive frames in the dataset
# x_batch, y_batch would be numpy arrays of size seq_length x maxNumPeds x 3
# d_batch would be a scalar identifying the dataset from which this sequence is extracted
x_batch, y_batch, d_batch = x[batch], y[batch], d[batch]
if d_batch == 0 and datasets[0] == 0:
dataset_data = [640, 480]
else:
dataset_data = [720, 576]
grid_batch = getSequenceGridMask(x_batch, dataset_data,
args.neighborhood_size,
args.grid_size)
# Feed the source, target data
feed = {
model.input_data: x_batch,
model.target_data: y_batch,
model.grid_data: grid_batch
}
train_loss = sess.run(model.cost, feed)
loss_batch += train_loss
loss_batch = loss_batch / data_loader.batch_size
loss_epoch += loss_batch
loss_epoch /= data_loader.valid_num_batches
# Update best validation loss until now
if loss_epoch < best_val_loss:
best_val_loss = loss_epoch
best_epoch = e
print('(epoch {}), valid_loss = {:.3f}'.format(e, loss_epoch))
print('Best epoch', best_epoch, 'Best validation loss',
best_val_loss)
log_file_curve.write(str(loss_epoch) + '\n')
print('*****************')
# Save the model after each epoch
print('Saving model')
checkpoint_path = os.path.join(save_directory, 'social_model.ckpt')
saver.save(sess, checkpoint_path, global_step=e)
print("model saved to {}".format(checkpoint_path))
print('Best epoch', best_epoch, 'Best validation loss', best_val_loss)
log_file.write(str(best_epoch) + ',' + str(best_val_loss))
# CLose logging files
log_file.close()
log_file_curve.close()
def main():
# Set random seed
np.random.seed(1)
parser = argparse.ArgumentParser()
# Observed length of the trajectory parameter
parser.add_argument('--obs_length',
type=int,
default=6,
help='Observed length of the trajectory')
# Predicted length of the trajectory parameter
parser.add_argument('--pred_length',
type=int,
default=6,
help='Predicted length of the trajectory')
# Test dataset
parser.add_argument('--test_dataset',
type=int,
default=3,
help='Dataset to be tested on')
# Model to be loaded
parser.add_argument('--epoch',
type=int,
default=0,
help='Epoch of model to be loaded')
# Parse the parameters
# sample_args = parser.parse_args()
args = parser.parse_args()
# Save directory
save_directory = 'save/' + str(args.test_dataset) + '/'
# Define the path for the config file for saved args
with open(os.path.join(save_directory, 'social_config.pkl'), 'rb') as f:
saved_args = pickle.load(f)
# Create a SocialModel object with the saved_args and infer set to true
model = SocialModel(saved_args, True)
# Initialize a TensorFlow session
sess = tf.InteractiveSession()
# Initialize a saver
saver = tf.train.Saver()
# Get the checkpoint state for the model
ckpt = tf.train.get_checkpoint_state(save_directory)
# print ('loading model: ', ckpt.model_checkpoint_path)
# print('hahah: ', len(ckpt.all_model_checkpoint_paths))
print('loading model: ', ckpt.all_model_checkpoint_paths[args.epoch])
# Restore the model at the checkpoint
saver.restore(sess, ckpt.all_model_checkpoint_paths[args.epoch])
# Dataset to get data from
dataset = [0]
# Create a SocialDataLoader object with batch_size 1 and seq_length equal to observed_length + pred_length
data_loader = SocialDataLoader(1,
args.pred_length + args.obs_length,
saved_args.maxNumPeds,
dataset,
True,
infer=True)
# Reset all pointers of the data_loader
data_loader.reset_batch_pointer()
results = []
# Variable to maintain total error
total_error = 0
# For each batch
for b in range(data_loader.num_batches):
# Get the source, target and dataset data for the next batch
x, y, d = data_loader.next_batch(randomUpdate=False)
# Batch size is 1
x_batch, y_batch, d_batch = x[0], y[0], d[0]
# if d_batch == 0 and dataset[0] == 0:
# dimensions = [640, 480]
# else:
# dimensions = [720, 576]
dimensions = [1640, 78]
grid_batch = getSequenceGridMask(x_batch, dimensions,
saved_args.neighborhood_size,
saved_args.grid_size)
obs_traj = x_batch[:args.obs_length]
obs_grid = grid_batch[:args.obs_length]
# obs_traj is an array of shape obs_length x maxNumPeds x 3
print "********************** SAMPLING A NEW TRAJECTORY", b, "******************************"
complete_traj = model.sample(sess, obs_traj, obs_grid, dimensions,
x_batch, args.pred_length)
# ipdb.set_trace()
# complete_traj is an array of shape (obs_length+pred_length) x maxNumPeds x 3
print('hahah', len(complete_traj))
total_error += get_mean_error(complete_traj, x[0], args.obs_length,
saved_args.maxNumPeds)
print "Processed trajectory number : ", b, "out of ", data_loader.num_batches, " trajectories"
# plot_trajectories(x[0], complete_traj, sample_args.obs_length)
# return
results.append((x[0], complete_traj, args.obs_length))
# Print the mean error across all the batches
print "Total mean error of the model is ", total_error / data_loader.num_batches
print "Saving results"
with open(os.path.join(save_directory, 'social_results.pkl'), 'wb') as f:
pickle.dump(results, f)
def train(args):
datasets = list(range(5))
# Remove the leaveDataset from data_sets
datasets.remove(args.test_dataset)
# Create the data loader object. This object would preprocess the data in terms of
# batches each of size args.batch_size, of length args.seq_length
data_loader = DataLoader(args.batch_size,
args.obs_length,
args.obs_length,
maxNumPeds=args.maxNumPeds,
datasets=datasets,
forcePreProcess=True)
# https://stackoverflow.com/a/41146954/2049763
import pathlib
# Log directory
log_directory = os.path.join(args.train_logs, 'log',
str(args.test_dataset))
path = pathlib.Path(log_directory)
path.mkdir(parents=True, exist_ok=True)
# Logging files
log_file_curve = open(os.path.join(log_directory, 'log_curve.txt'), 'w')
log_file = open(os.path.join(log_directory, 'val.txt'), 'w')
# Save directory
save_directory = os.path.join(args.train_logs, 'save',
str(args.test_dataset))
path = pathlib.Path(save_directory)
path.mkdir(parents=True, exist_ok=True)
# model directory
model_directory = os.path.join(args.train_logs, 'model',
str(args.test_dataset))
path = pathlib.Path(model_directory)
path.mkdir(parents=True, exist_ok=True)
# Save the arguments int the config file
with open(os.path.join(save_directory, 'config.pkl'), 'wb') as f:
pickle.dump(args, f)
checkpoint_path = os.path.join(model_directory, 'model.ckpt')
# Create a Vanilla LSTM model with the arguments
model = Model(args)
config = tf.ConfigProto()
config.gpu_options.allow_growth = True
# Showing which device is allocated (in case of multiple GPUs)
config = tf.ConfigProto(log_device_placement=True)
# Allocating 70% of memory in each GPU with 0.5
config.gpu_options.per_process_gpu_memory_fraction = 0.7
# Initialize a TensorFlow session
with tf.Session() as sess:
# Summaries need to be displayed
# Whenever you need to record the loss, feed the mean loss to this placeholder
tf_loss_ph = tf.placeholder(tf.float32,
shape=None,
name='loss_summary')
# Create a scalar summary object for the loss so it can be displayed
tf_loss_summary = tf.summary.scalar('loss', tf_loss_ph)
# Whenever you need to record the loss, feed the mean loss to this placeholder
tf_embedding_w_ph = tf.placeholder(tf.float32,
shape=None,
name='embedding_w_summary')
tf_embedding_w_summary = tf.summary.scalar('embedding_w',
tf_embedding_w_ph)
# Whenever you need to record the loss, feed the mean loss to this placeholder
tf_output_w_ph = tf.placeholder(tf.float32,
shape=None,
name='output_w_summary')
tf_output_w_summary = tf.summary.scalar('output_w', tf_output_w_ph)
# Whenever you need to record the loss, feed the mean loss to this placeholder
tf_lr_ph = tf.placeholder(tf.float32,
shape=None,
name='learning_rate_summary')
tf_lr_ph_summary = tf.summary.scalar('learning_rate', tf_lr_ph)
# Whenever you need to record the loss, feed the mean loss to this placeholder
# tf_val_loss_ph = tf.placeholder(tf.float32, shape=None, name='val_loss_summary')
# tf_val_loss_summary = tf.summary.scalar('val_loss', tf_val_loss_ph)
# Whenever you need to record the loss, feed the mean loss to this placeholder
tf_val_error_ph = tf.placeholder(tf.float32,
shape=None,
name='val_error_summary')
tf_val_error_summary = tf.summary.scalar('val_error', tf_val_error_ph)
# https://stackoverflow.com/a/40148954/2049763
train_writer = tf.summary.FileWriter(model_directory, sess.graph)
val_writer = tf.summary.FileWriter(
os.path.join(model_directory, 'eval'))
# Initialize all the variables in the graph
sess.run(tf.global_variables_initializer())
# Add all the variables to the list of variables to be saved
saver = tf.train.Saver(tf.global_variables())
best_val_loss = 100
best_epoch = 0
print("**** Training is starting !")
# For each epoch
for epoch in range(args.num_epochs):
# Assign the learning rate (decayed acc. to the epoch number)
# learning_rate = args.learning_rate * (args.decay_rate ** epoch)
# Reset the pointers in the data loader object
data_loader.reset_batch_pointer()
loss_per_epoch = []
total_steps = args.num_epochs * data_loader.num_batches
# For each batch in this epoch
for batch in range(data_loader.num_batches):
# Tic
start = time.time()
# Get the source and target data of the current batch
# x has the source data, y has the target data
x, y, d = data_loader.next_batch(randomUpdate=True)
# variable to store the loss for this batch
loss_per_batch = []
gradients = {}
# For each sequence in the batch
for sequence in range(data_loader.batch_size):
# x_batch, y_batch and d_batch contains the source, target and dataset index data for
# seq_length long consecutive frames in the dataset
# x_batch, y_batch would be numpy arrays of size seq_length x maxNumPeds x 3
# d_batch would be a scalar identifying the dataset from which this sequence is extracted
x_batch, y_batch = x[sequence], y[sequence]
# Feed the source, target data
feed = {
model.input_data: x_batch,
model.target_data: y_batch,
model.keep_prob: args.keep_prob,
model.lr: args.learning_rate,
model.training_epoch: epoch
}
# train_loss, gradient, _, lr = sess.run(
# [model.cost, model.clipped_gradients, model.train_op, model.final_lr], feed)
train_loss, gradient, lr = sess.run(
[model.cost, model.clipped_gradients, model.final_lr],
feed)
vars = tf.trainable_variables()
# vars_vals = sess.run(vars)
for var, val in zip(vars, gradient):
# print(var.name, np.shape(val))
if var.name in gradients:
gradients[var.name].append(val)
else:
gradients[var.name] = []
if not np.isnan(train_loss):
loss_per_batch.append(train_loss)
# break
else:
print(
"epoch#{} batch#{} sequence#{} train_loss is NaN".
format(epoch + 1, batch + 1, sequence))
avg_loss_per_batch = np.mean(loss_per_batch)
loss_per_epoch.append(avg_loss_per_batch)
my_global_step = (epoch + 1) * data_loader.num_batches + batch
# Print epoch, batch, loss and time taken
print(
"## {}/{} (Epoch: {}/{}), (Batch: {}/{}) train_loss = {:.3f}, time/batch = {:.3f}"
.format(my_global_step, total_steps, epoch + 1,
args.num_epochs, batch + 1,
data_loader.num_batches, avg_loss_per_batch,
time.time() - start))
# Save the model if the current epoch and batch number match the frequency
if my_global_step % args.save_every == 0 and (my_global_step >
0):
saver.save(sess,
checkpoint_path,
global_step=my_global_step)
print("model saved to {}".format(checkpoint_path))
# break
# ######## **** Mini batch optimization starts **** ########
# vars_vals = sess.run(vars)
feed, grad_ph_key = {
model.lr: args.learning_rate,
model.training_epoch: epoch
}, 0
for var, val_ in gradients.items():
val = np.mean(val_, axis=0)
feed[model.grad_placeholders[grad_ph_key]] = val
grad_ph_key += 1
# print(var, np.shape(val_), np.shape(val))
val = np.sum(np.absolute(val))
# print("var: {}, value: {}".format(var, val))
if 'embedding_w' in var:
embedding_w_summary = val
elif 'output_w' in var:
output_w_summary = val
_ = sess.run([model.train_op_2], feed)
# ######## **** Mini batch optimization ends **** ########
# ######## **** Training batch iteration ends **** ########
avg_loss_per_epoch = np.mean(loss_per_epoch)
print('# (Epoch {}/{}), Learning rate = {} Training Loss = {:.3f}'.
format(epoch + 1, args.num_epochs, lr, avg_loss_per_epoch))
log_file_curve.write(
str(epoch) + ',' + str(avg_loss_per_epoch) + ',')
# Execute the summaries defined above
training_loss_summary, embedding_w_summary, output_w_summary, lr_ph_summary = sess.run(
[
tf_loss_summary, tf_embedding_w_summary,
tf_output_w_summary, tf_lr_ph_summary
],
feed_dict={
tf_loss_ph: avg_loss_per_epoch,
tf_embedding_w_ph: embedding_w_summary,
tf_output_w_ph: output_w_summary,
tf_lr_ph: lr
})
training_summaries = tf.summary.merge([
training_loss_summary, embedding_w_summary, output_w_summary,
lr_ph_summary
])
training_summaries_tensor = sess.run(training_summaries)
# ######## **** Validation starts **** ########
data_loader.reset_batch_pointer(valid=True)
val_loss_per_epoch, val_error_per_epoch = [], []
for batch in range(data_loader.num_batches):
# Get the source, target and dataset data for the next batch x, y are input and target data which are
# lists containing numpy arrays of size seq_length x maxNumPeds x 3 d is the list of dataset indices
# from which each batch is generated (used to differentiate between datasets)
x, y, d = data_loader.next_batch(valid=True, randomUpdate=True)
# variable to store the loss for this batch
val_loss_per_batch, val_error_per_batch = [], []
# For each sequence in the batch
for sequence in range(data_loader.batch_size):
# x_batch, y_batch and d_batch contains the source, target and dataset index data for
# seq_length long consecutive frames in the dataset
# x_batch, y_batch would be numpy arrays of size seq_length x maxNumPeds x 3
# d_batch would be a scalar identifying the dataset from which this sequence is extracted
x_batch, y_batch = x[sequence], y[sequence]
# Feed the source, target data
feed = {
model.input_data: x_batch,
model.target_data: y_batch,
model.keep_prob: 1.
}
output, val_loss = sess.run(
[model.final_result, model.cost], feed)
val_loss_per_batch.append(val_loss)
val_error_per_batch.append(
model.training_mean_error(x_batch, y_batch, output))
val_loss_per_epoch.append(np.mean(val_loss_per_batch))
val_error_per_epoch.append(np.mean(val_error_per_batch))
# break
avg_val_loss_per_epoch = np.mean(val_loss_per_epoch)
avg_val_error_per_epoch = np.mean(val_error_per_epoch)
# Update best validation loss until now
if avg_val_loss_per_epoch < best_val_loss:
best_val_loss = avg_val_loss_per_epoch
best_epoch = epoch
print('# (Epoch {}/{}), Validation loss = {:.3f}, error = {:.3f}'.
format(epoch + 1, args.num_epochs, avg_val_loss_per_epoch,
avg_val_error_per_epoch))
log_file_curve.write(str(avg_val_loss_per_epoch) + '\n')
# Execute the summaries defined above
val_loss_summary, val_error_summary = sess.run(
[tf_loss_summary, tf_val_error_summary],
feed_dict={
tf_loss_ph: avg_val_loss_per_epoch,
tf_val_error_ph: avg_val_error_per_epoch
})
# Merge all summaries together
performance_summaries = tf.summary.merge(
[val_loss_summary, val_error_summary])
# https://stackoverflow.com/a/51784126/2049763
performance_summaries_tensor = sess.run(performance_summaries)
# Write the obtained summaries to the file, so it can be displayed in the TensorBoard
train_writer.add_summary(training_summaries_tensor, epoch)
val_writer.add_summary(performance_summaries_tensor, epoch)
print('Best epoch', best_epoch, 'Best validation loss', best_val_loss)
log_file.write(str(best_epoch) + ',' + str(best_val_loss))
my_global_step += 1
saver.save(sess, checkpoint_path, global_step=my_global_step)
print("model saved to {}".format(checkpoint_path))
# CLose logging files
log_file.close()
log_file_curve.close()
train_writer.close()
val_writer.close()
def train(args):
if args.visible:
os.environ["CUDA_VISIBLE_DEVICES"] = args.visible
save_path = make_save_path(args)
dataset_path = args.dataset_path
log_path = os.path.join(save_path, 'log')
if not os.path.isdir(log_path):
os.makedirs(log_path)
# Create the SocialDataLoader object
data_loader = SocialDataLoader(args.batch_size, args.seq_length,
args.maxNumPeds, dataset_path, forcePreProcess=True)
with open(os.path.join(save_path, 'social_config.pkl'), 'wb') as f:
pickle.dump(args, f)
# Create a SocialModel object with the arguments
model = SocialModel(args)
all_loss = []
# Initialize a TensorFlow session
with tf.Session() as sess:
# Initialize all variables in the graph
sess.run(tf.initialize_all_variables())
# Initialize a saver that saves all the variables in the graph
saver = tf.train.Saver(tf.all_variables())
summary_writer = tf.summary.FileWriter(log_path, sess.graph)
# For each epoch
for e in range(args.num_epochs):
# Assign the learning rate value for this epoch
sess.run(tf.assign(model.lr, args.learning_rate * (args.decay_rate ** e)))
# Reset the data pointers in the data_loader
data_loader.reset_batch_pointer()
# For each batch
for b in range(data_loader.num_batches):
# Tic
start = time.time()
# Get the source, target and dataset data for the next batch
# s_batch, t_batch are input and target data which are lists containing numpy arrays of size seq_length x maxNumPeds x 3
# d is the list of dataset indices from which each batch is generated (used to differentiate between datasets)
s_batch, t_batch, d = data_loader.next_batch()
# variable to store the loss for this batch
loss_batch = 0
# For each sequence in the batch
for seq_num in range(data_loader.batch_size):
# s_seq, t_seq and d_batch contains the source, target and dataset index data for
# seq_length long consecutive frames in the dataset
# s_seq, t_seq would be numpy arrays of size seq_length x maxNumPeds x 3
# d_batch would be a scalar identifying the dataset from which this sequence is extracted
s_seq, t_seq, d_seq = s_batch[seq_num], t_batch[seq_num], d[seq_num]
'''
if d_seq == 0 and datasets[0] == 0:
dataset_data = [640, 480]
else:
dataset_data = [720, 576]
'''
grid_batch = getSequenceGridMask(s_seq, [0, 0], args.neighborhood_size, args.grid_size)
# Feed the source, target data
feed = {model.input_data: s_seq, model.target_data: t_seq, model.grid_data: grid_batch}
train_loss, _ = sess.run([model.cost, model.train_op], feed)
loss_batch += train_loss
end = time.time()
loss_batch = loss_batch / data_loader.batch_size
all_loss.append(loss_batch)
print(
"{}/{} (epoch {}), train_loss = {:.3f}, time/seq_num = {:.3f}"
.format(
e * data_loader.num_batches + b,
args.num_epochs * data_loader.num_batches,
e,
loss_batch, end - start))
# Save the model if the current epoch and batch number match the frequency
if (e * data_loader.num_batches + b) % args.save_every == 0 and ((e * data_loader.num_batches + b) > 0):
checkpoint_path = os.path.join(save_path, 'social_model.ckpt')
saver.save(sess, checkpoint_path, global_step=e * data_loader.num_batches + b)
print("model saved to {}".format(checkpoint_path))
np.savetxt(os.path.join(log_path, 'loss.txt'), np.asarray(all_loss))
def main():
# Set random seed
np.random.seed(1)
parser = argparse.ArgumentParser()
# Observed length of the trajectory parameter
parser.add_argument('--obs_length', type=int, default=8,
help='Observed length of the trajectory')
# Predicted length of the trajectory parameter
parser.add_argument('--pred_length', type=int, default=12,
help='Predicted length of the trajectory')
# Test dataset
parser.add_argument('--test_dataset', type=str,
help='Dataset to be tested on')
parser.add_argument('--visible',type=str,
required=False, default=None, help='GPU to run on')
parser.add_argument('--model_path', type=str)
# Parse the parameters
sample_args = parser.parse_args()
if sample_args.visible:
os.environ["CUDA_VISIBLE_DEVICES"] = sample_args.visible
save_path = sample_args.model_path
# Define the path for the config file for saved args
with open(os.path.join(save_path, 'social_config.pkl'), 'rb') as f:
saved_args = pickle.load(f)
# Create a SocialModel object with the saved_args and infer set to true
model = SocialModel(saved_args, True)
# Initialize a TensorFlow session
sess = tf.InteractiveSession()
# Initialize a saver
saver = tf.train.Saver()
# Get the checkpoint state for the model
ckpt = tf.train.get_checkpoint_state(save_path)
print ('loading model: ', ckpt.model_checkpoint_path)
# Restore the model at the checkpoint
saver.restore(sess, ckpt.model_checkpoint_path)
# Create a SocialDataLoader object with batch_size 1 and seq_length equal to observed_length + pred_length
data_loader = SocialDataLoader(1, sample_args.pred_length +
sample_args.obs_length, saved_args.maxNumPeds, sample_args.test_dataset, True)
# Reset all pointers of the data_loader
data_loader.reset_batch_pointer()
results = []
# Variable to maintain total error
total_error = 0
# For each batch
for b in range(data_loader.num_batches):
# Get the source, target and dataset data for the next batch
x, y, d = data_loader.next_batch(randomUpdate=False)
# Batch size is 1
x_batch, y_batch, d_batch = x[0], y[0], d[0]
'''
if d_batch == 0 and dataset[0] == 0:
dimensions = [640, 480]
else:
dimensions = [720, 576]
'''
grid_batch = getSequenceGridMask(x_batch, [0,0], saved_args.neighborhood_size, saved_args.grid_size)
obs_traj = x_batch[:sample_args.obs_length]
obs_grid = grid_batch[:sample_args.obs_length]
# obs_traj is an array of shape obs_length x maxNumPeds x 3
print "********************** SAMPLING A NEW TRAJECTORY", b, "******************************"
complete_traj = model.sample(sess, obs_traj, obs_grid, [0,0], x_batch, sample_args.pred_length)
# ipdb.set_trace()
# complete_traj is an array of shape (obs_length+pred_length) x maxNumPeds x 3
total_error += get_mean_error(complete_traj, x[0], sample_args.obs_length, saved_args.maxNumPeds)
print "Processed trajectory number : ", b, "out of ", data_loader.num_batches, " trajectories"
# plot_trajectories(x[0], complete_traj, sample_args.obs_length)
# return
results.append((x[0], complete_traj, sample_args.obs_length))
# Print the mean error across all the batches
print "Total mean error of the model is ", total_error/data_loader.num_batches
print "Saving results"
with open(os.path.join(save_path, 'social_results.pkl'), 'wb') as f:
pickle.dump(results, f)
with open(os.path.join(save_directory, 'social_config.pkl'), 'rb') as f:
saved_args = pickle.load(f)
#f = open('/home/hesl/PycharmProjects/social-lstm-tf-HW/ResultofTrainingKITTI-13NTestonKITTI-17/save/social_results.pkl', 'rb')
#f = open('/home/hesl/PycharmProjects/social-lstm-tf-HW/ResultofTrainingETH1TestETH0/save/social_results.pkl', 'rb')
f = open(
'/home/hesl/PycharmProjects/social-lstm-tf-HW/save/social_results.pkl',
'rb')
results = pickle.load(f)
dataset = [sample_args.visual_dataset]
data_loader = SocialDataLoader(1,
sample_args.pred_length +
sample_args.obs_length,
saved_args.maxNumPeds,
dataset,
True,
infer=True)
#[7,16]
#print(data_loader.data[0][0].shape)
#
# '''Visualize Ground Truth (u,v)'''
# for j in range(len(data_loader.frameList[0])):
#
# #sourceFileName = "/home/hesl/PycharmProjects/social-lstm-tf-HW/data/KITTI-17/img1/" + str(j + 1).zfill(6) + ".jpg"
# #Visualize ETH/hotel
# #sourceFileName = "/media/hesl/OS/Documents and Settings/N1701420F/Desktop/video/eth/hotel/frame-" + str(int(data_loader.frameList[0][j])).zfill(3) + ".jpg"
# #Eth/eth
# #sourceFileName = "/media/hesl/OS/Documents and Settings/N1701420F/Desktop/video/eth/eth/frame-" + str(int(data_loader.frameList[0][j])).zfill(3)+ ".jpeg"
def sample(args, save_location, model_directory):
results_pkl = os.path.join(save_location, 'results.pkl')
with open(os.path.join(save_location, 'config.pkl'), 'rb') as f:
saved_args = pickle.load(f)
# Create a SocialModel object with the saved_args and infer set to true
model = Model(saved_args, True)
# Initialize a TensorFlow session
config = tf.ConfigProto(
log_device_placement=True
) # Showing which device is allocated (in case of multiple GPUs)
config.gpu_options.per_process_gpu_memory_fraction = 0.4 # Allocating 40% of memory in each GPU
sess = tf.InteractiveSession(config=config)
ckpt = tf.train.get_checkpoint_state(model_directory)
# Initialize a saver
saver = tf.train.Saver()
# Restore the model at the checkpoint
saver.restore(sess, ckpt.model_checkpoint_path)
# Dataset to get data from
dataset = [args.test_dataset]
# Create a SocialDataLoader object with batch_size 1 and seq_length equal to observed_length + pred_length
data_loader = SocialDataLoader(1,
args.obs_length,
args.pred_length,
saved_args.maxNumPeds,
dataset,
forcePreProcess=True,
infer=True)
# Reset all pointers of the data_loader
data_loader.reset_batch_pointer()
results = []
# Variable to maintain total error
total_error = 0
final_displacement_error = []
# For each batch
for b in range(data_loader.num_batches):
# Get the source, target and dataset data for the next batch
x, y, d = data_loader.next_batch(randomUpdate=False)
# Batch size is 1
x_batch, y_batch = x[0], y[0]
true_traj = np.concatenate((x_batch, y_batch[-args.pred_length:]),
axis=0)
# complete_traj is an array of shape ( obs_length + pred_length ) x maxNumPeds x 3
complete_traj = model.sample(sess, x_batch, true_traj,
args.pred_length)
total_error += model.get_mean_error(complete_traj, true_traj,
args.obs_length,
saved_args.maxNumPeds)
final_error = model.get_final_displacement_error(
complete_traj, true_traj, saved_args.maxNumPeds)
if final_error is not None:
final_displacement_error.append(final_error)
print("Processed trajectory number : ", b, "out of ",
data_loader.num_batches, " trajectories")
results.append((true_traj, complete_traj, args.obs_length))
print("Saving results")
with open(results_pkl, 'wb') as f:
pickle.dump(results, f)
# Print the mean error across all the batches
print("Total mean error of the model is {:.3f}".format(
total_error / data_loader.num_batches))
print("Total final error of the model is {:.3f}".format(
np.mean(final_displacement_error)))
def main():
np.random.seed(1)
parser = argparse.ArgumentParser()
# 观测轨迹长度
parser.add_argument('--obs_length',
type=int,
default=7,
help='Observed length of the trajectory')
# 预测轨迹长度
parser.add_argument('--pred_length',
type=int,
default=5,
help='Predicted length of the trajectory')
# 测试数据集
parser.add_argument('--test_dataset',
type=int,
default=2,
help='Epoch of model to be loaded')
# 导入的模型
parser.add_argument('--epoch',
type=int,
default=8,
help='Epoch of model to be loaded')
sample_args = parser.parse_args(args=[])
# 存储历史
save_directory = 'save/' + str(sample_args.test_dataset) + '/'
# Define the path for the config file for saved args
with open(os.path.join(save_directory, 'social_config.pkl'), 'rb') as f:
saved_args = pickle.load(f)
# Create a SocialModel object with the saved_args and infer set to true
model = SocialModel(saved_args, True)
# Initialize a TensorFlow session
config = tf.ConfigProto(
log_device_placement=True
) # Showing which device is allocated (in case of multiple GPUs)
config.gpu_options.per_process_gpu_memory_fraction = 0.8 # Allocating 20% of memory in each GPU
sess = tf.InteractiveSession(config=config)
# Initialize a saver
saver = tf.train.Saver()
# Get the checkpoint state for the model
ckpt = tf.train.get_checkpoint_state(save_directory)
# print ('loading model: ', ckpt.model_checkpoint_path)
print('loading model: ',
ckpt.all_model_checkpoint_paths[sample_args.epoch])
# Restore the model at the checkpoint
saver.restore(sess, ckpt.all_model_checkpoint_paths[sample_args.epoch])
# Dataset to get data from
dataset = [sample_args.test_dataset]
# Create a SocialDataLoader object with batch_size 1 and seq_length equal to observed_length + pred_length
data_loader = SocialDataLoader(1,
sample_args.pred_length +
sample_args.obs_length,
saved_args.maxNumPeds,
dataset,
True,
infer=True)
# Reset all pointers of the data_loader
data_loader.reset_batch_pointer()
results = []
# Variable to maintain total error
total_error = 0
# For each batch
for b in range(data_loader.num_batches):
# Get the source, target and dataset data for the next batch
x, y, d = data_loader.next_batch(randomUpdate=False)
# Batch size is 1
x_batch, y_batch, d_batch = x[0], y[0], d[0]
if d_batch == 0 and dataset[0] == 0:
dimensions = [640, 480]
else:
dimensions = [720, 576]
grid_batch = getSequenceGridMask(x_batch, dimensions,
saved_args.neighborhood_size,
saved_args.grid_size)
obs_traj = x_batch[:sample_args.obs_length]
obs_grid = grid_batch[:sample_args.obs_length]
# obs_traj is an array of shape obs_length x maxNumPeds x 3
print("********************** SAMPLING A NEW TRAJECTORY", b,
"******************************")
complete_traj = model.sample(sess, obs_traj, obs_grid, dimensions,
x_batch, sample_args.pred_length)
# ipdb.set_trace()
# complete_traj is an array of shape (obs_length+pred_length) x maxNumPeds x 3
total_error += get_mean_error(complete_traj, x[0],
sample_args.obs_length,
saved_args.maxNumPeds)
print("Processed trajectory number : ", b, "out of ",
data_loader.num_batches, " trajectories")
print('Model loaded: ',
ckpt.all_model_checkpoint_paths[sample_args.epoch])
# plot_trajectories(x[0], complete_traj, sample_args.obs_length)
# return
results.append((x[0], complete_traj, sample_args.obs_length))
# Print the mean error across all the batches
print("Total mean error of the model is ",
total_error / data_loader.num_batches)
print("Saving results")
with open(os.path.join(save_directory, 'social_results.pkl'), 'wb') as f:
pickle.dump(results, f)