def main():
db = vmdatabase.VMDatabase()
machine = db.getMachineByJenkinsName(NODE_NAME)
if machine.state != vmdatabase.HOLD:
machine.state = vmdatabase.DELETE
try:
utils.update_stats(machine.base_image.provider)
if UPSTREAM_BUILD_URL:
fd = urllib.urlopen(UPSTREAM_BUILD_URL+'api/json')
data = json.load(fd)
result = data['result']
if statsd and result == 'SUCCESS':
dt = int(data['duration'])
key = 'devstack.job.%s' % UPSTREAM_JOB_NAME
statsd.timing(key+'.runtime', dt)
statsd.incr(key+'.builds')
key += '.%s' % UPSTREAM_BRANCH
statsd.timing(key+'.runtime', dt)
statsd.incr(key+'.builds')
key += '.%s' % machine.base_image.provider.name
statsd.timing(key+'.runtime', dt)
statsd.incr(key+'.builds')
except:
print "Error getting build information"
traceback.print_exc()
def main():
db = vmdatabase.VMDatabase()
if not SKIP_DEVSTACK_GATE_JENKINS:
config = ConfigParser.ConfigParser()
config.read(DEVSTACK_GATE_SECURE_CONFIG)
jenkins = myjenkins.Jenkins(config.get('jenkins', 'server'),
config.get('jenkins', 'user'),
config.get('jenkins', 'apikey'))
jenkins.get_info()
else:
jenkins = None
provider = db.getProvider(PROVIDER_NAME)
print "Working with provider %s" % provider.name
error = False
for machine in provider.machines:
if machine.state != vmdatabase.READY:
continue
print 'Checking machine', machine.name
try:
check_machine(jenkins, machine)
except:
error = True
traceback.print_exc()
utils.update_stats(provider)
if error:
sys.exit(1)
def main():
db = vmdatabase.VMDatabase()
if not SKIP_DEVSTACK_GATE_JENKINS:
config = ConfigParser.ConfigParser()
config.read(DEVSTACK_GATE_SECURE_CONFIG)
jenkins = myjenkins.Jenkins(config.get('jenkins', 'server'),
config.get('jenkins', 'user'),
config.get('jenkins', 'apikey'))
jenkins.get_info()
else:
jenkins = None
provider = db.getProvider(PROVIDER_NAME)
print "Working with provider %s" % provider.name
error = False
for machine in provider.machines:
if machine.state != vmdatabase.READY:
continue
print 'Checking machine', machine.name
try:
check_machine(jenkins, machine)
except:
error = True
traceback.print_exc()
utils.update_stats(provider)
if error:
sys.exit(1)
def main():
db = vmdatabase.VMDatabase()
machine = db.getMachineByJenkinsName(NODE_NAME)
if machine.state != vmdatabase.HOLD:
machine.state = vmdatabase.DELETE
utils.update_stats(machine.base_image.provider)
def main():
db = vmdatabase.VMDatabase()
config = ConfigParser.ConfigParser()
config.read(DEVSTACK_GATE_SECURE_CONFIG)
if not SKIP_DEVSTACK_GATE_JENKINS:
jenkins = myjenkins.Jenkins(config.get('jenkins', 'server'),
config.get('jenkins', 'user'),
config.get('jenkins', 'apikey'))
jenkins.get_info()
else:
jenkins = None
try:
machine = db.getMachineByJenkinsName(NODE_NAME)
except Exception:
utils.log.debug("Unable to find node: %s" % NODE_NAME)
return
utils.log.debug("Used ID: %s old state: %s build:%s" % (
machine.id, machine.state, BUILD_URL))
machine.state = vmdatabase.USED
if jenkins:
if machine.jenkins_name:
if jenkins.node_exists(machine.jenkins_name):
config = jenkins.get_node_config(machine.jenkins_name)
old = None
m = LABEL_RE.search(config)
if m:
old = m.group(1)
config = LABEL_RE.sub('<label>devstack-used</label>', config)
for i in range(3):
try:
jenkins.reconfig_node(machine.jenkins_name, config)
except:
if i == 2:
utils.log.exception(
"Unable to relabel ID: %s" % machine.id)
raise
time.sleep(5)
utils.log.debug(
"Relabeled ID: %s old label: %s new label: %s" % (
machine.id, old, 'devstack-used'))
utils.update_stats(machine.base_image.provider)
def main():
db = vmdatabase.VMDatabase()
config = ConfigParser.ConfigParser()
config.read(DEVSTACK_GATE_SECURE_CONFIG)
if not SKIP_DEVSTACK_GATE_JENKINS:
jenkins = myjenkins.Jenkins(config.get('jenkins', 'server'),
config.get('jenkins', 'user'),
config.get('jenkins', 'apikey'))
jenkins.get_info()
else:
jenkins = None
try:
machine = db.getMachineByJenkinsName(NODE_NAME)
except Exception:
utils.log.debug("Unable to find node: %s" % NODE_NAME)
return
utils.log.debug("Used ID: %s old state: %s build:%s" %
(machine.id, machine.state, BUILD_URL))
machine.state = vmdatabase.USED
if jenkins:
if machine.jenkins_name:
if jenkins.node_exists(machine.jenkins_name):
config = jenkins.get_node_config(machine.jenkins_name)
old = None
m = LABEL_RE.search(config)
if m:
old = m.group(1)
config = LABEL_RE.sub('<label>devstack-used</label>', config)
for i in range(3):
try:
jenkins.reconfig_node(machine.jenkins_name, config)
except:
if i == 2:
utils.log.exception("Unable to relabel ID: %s" %
machine.id)
raise
time.sleep(5)
utils.log.debug(
"Relabeled ID: %s old label: %s new label: %s" %
(machine.id, old, 'devstack-used'))
utils.update_stats(machine.base_image.provider)
def main():
db = vmdatabase.VMDatabase()
try:
machine = db.getMachineByJenkinsName(NODE_NAME)
except Exception:
utils.log.debug("Unable to find node: %s" % NODE_NAME)
return
if machine.state != vmdatabase.HOLD:
utils.log.debug("Set deleted ID: %s old state: %s build: %s" % (
machine.id, machine.state, BUILD_URL))
machine.state = vmdatabase.DELETE
else:
utils.log.debug("Hold ID: %s old state: %s build: %s" % (
machine.id, machine.state, BUILD_URL))
try:
utils.update_stats(machine.base_image.provider)
if UPSTREAM_BUILD_URL:
fd = urllib.urlopen(UPSTREAM_BUILD_URL + 'api/json')
data = json.load(fd)
result = data['result']
if statsd and result == 'SUCCESS':
dt = int(data['duration'])
key = 'devstack.job.%s' % UPSTREAM_JOB_NAME
statsd.timing(key + '.runtime', dt)
statsd.incr(key + '.builds')
key += '.%s' % UPSTREAM_BRANCH
statsd.timing(key + '.runtime', dt)
statsd.incr(key + '.builds')
key += '.%s' % machine.base_image.provider.name
statsd.timing(key + '.runtime', dt)
statsd.incr(key + '.builds')
except:
print "Error getting build information"
traceback.print_exc()
def main():
db = vmdatabase.VMDatabase()
config = ConfigParser.ConfigParser()
config.read(DEVSTACK_GATE_SECURE_CONFIG)
jenkins = myjenkins.Jenkins(config.get('jenkins', 'server'),
config.get('jenkins', 'user'),
config.get('jenkins', 'apikey'))
jenkins.get_info()
machine = db.getMachineByJenkinsName(NODE_NAME)
machine.state = vmdatabase.USED
if machine.jenkins_name:
if jenkins.node_exists(machine.jenkins_name):
config = jenkins.get_node_config(machine.jenkins_name)
config = LABEL_RE.sub('<label>devstack-used</label>', config)
jenkins.reconfig_node(machine.jenkins_name, config)
utils.update_stats(machine.base_image.provider)
def main():
db = vmdatabase.VMDatabase()
if not SKIP_DEVSTACK_GATE_JENKINS:
config = ConfigParser.ConfigParser()
config.read(DEVSTACK_GATE_SECURE_CONFIG)
jenkins = myjenkins.Jenkins(config.get('jenkins', 'server'),
config.get('jenkins', 'user'),
config.get('jenkins', 'apikey'))
jenkins.get_info()
else:
jenkins = None
print 'Known machines (start):'
db.print_state()
provider = db.getProvider(PROVIDER_NAME)
print "Working with provider %s" % provider.name
client = utils.get_client(provider)
flavor = utils.get_flavor(client, 1024)
print "Found flavor", flavor
error = False
now = time.time()
for machine in provider.machines:
# Normally, reap machines that have sat in their current state
# for 24 hours, unless that state is READY.
if (REAP_ALL_SERVERS or
(machine.state != vmdatabase.READY and
now - machine.state_time > MACHINE_LIFETIME) or
machine.state == vmdatabase.DELETE or
machine.state == vmdatabase.ERROR):
print 'Deleting machine', machine.name
try:
delete_machine(jenkins, client, machine)
except:
error = True
traceback.print_exc()
provider_min_ready = 0
for base_image in provider.base_images:
provider_min_ready += base_image.min_ready
for snap_image in base_image.snapshot_images:
# Normally, reap images that have sat in their current state
# for 24 hours, unless the image is the current snapshot
if (REAP_ALL_IMAGES or
(snap_image != base_image.current_snapshot and
now - snap_image.state_time > MACHINE_LIFETIME)):
print 'Deleting image', snap_image.name
try:
delete_image(client, snap_image)
except:
error = True
traceback.print_exc()
print
print 'Known machines (end):'
db.print_state()
utils.update_stats(provider)
if error:
sys.exit(1)
def main():
db = vmdatabase.VMDatabase()
if not SKIP_DEVSTACK_GATE_JENKINS:
config = ConfigParser.ConfigParser()
config.read(DEVSTACK_GATE_SECURE_CONFIG)
jenkins = myjenkins.Jenkins(config.get('jenkins', 'server'),
config.get('jenkins', 'user'),
config.get('jenkins', 'apikey'))
jenkins.get_info()
else:
jenkins = None
provider = db.getProvider(PROVIDER_NAME)
print "Working with provider %s" % provider.name
client = utils.get_client(provider)
last_name = ''
error_counts = {}
error = False
for base_image in provider.base_images:
snap_image = base_image.current_snapshot
if not snap_image:
continue
print "Working on image %s" % snap_image.name
flavor = utils.get_flavor(client, base_image.min_ram)
print "Found flavor", flavor
remote_snap_image = client.images.get(snap_image.external_id)
print "Found image", remote_snap_image
num_to_launch = calculate_deficit(provider, base_image)
for i in range(num_to_launch):
try:
server, machine = launch_node(client,
snap_image,
remote_snap_image,
flavor,
last_name)
last_name = machine.name
except:
traceback.print_exc()
error = True
while True:
utils.update_stats(provider)
building_machines = provider.building_machines
if not building_machines:
print "No more machines are building, finished."
break
print "Waiting on %s machines" % len(building_machines)
for machine in building_machines:
try:
check_machine(jenkins, client, machine, error_counts)
except:
traceback.print_exc()
print "Abandoning machine %s" % machine.id
machine.state = vmdatabase.ERROR
error = True
db.commit()
time.sleep(3)
if error:
sys.exit(1)
def main():
db = vmdatabase.VMDatabase()
jenkins = None
credentials_id = None
if not SKIP_DEVSTACK_GATE_JENKINS:
config = ConfigParser.ConfigParser()
config.read(DEVSTACK_GATE_SECURE_CONFIG)
jenkins = myjenkins.Jenkins(config.get('jenkins', 'server'),
config.get('jenkins', 'user'),
config.get('jenkins', 'apikey'))
jenkins.get_info()
if config.has_option('jenkins', 'credentials_id'):
credentials_id = config.get('jenkins', 'credentials_id')
provider = db.getProvider(PROVIDER_NAME)
print "Working with provider %s" % provider.name
client = utils.get_client(provider)
last_name = ''
error_counts = {}
error = False
for base_image in provider.base_images:
snap_image = base_image.current_snapshot
if not snap_image:
continue
print "Working on image %s" % snap_image.name
flavor = utils.get_flavor(client, base_image.min_ram)
print "Found flavor", flavor
remote_snap_image = client.images.get(snap_image.external_id)
print "Found image", remote_snap_image
num_to_launch = calculate_deficit(provider, base_image)
for i in range(num_to_launch):
try:
server, machine = launch_node(client, snap_image,
remote_snap_image, flavor,
last_name)
last_name = machine.name
except:
traceback.print_exc()
error = True
while True:
utils.update_stats(provider)
building_machines = provider.building_machines
if not building_machines:
print "No more machines are building, finished."
break
print "Waiting on %s machines" % len(building_machines)
for machine in building_machines:
try:
check_machine(jenkins, client, machine, error_counts,
credentials_id)
except:
traceback.print_exc()
print "Abandoning machine %s" % machine.id
utils.log.exception("Abandoning ID: %s" % machine.id)
machine.state = vmdatabase.ERROR
error = True
db.commit()
time.sleep(3)
if error:
sys.exit(1)
def run(opt):
train_loader, test_loader = create_loaders(opt)
# Initialize generator and discriminator
generator = load_or_init_models(RetouchGenerator(opt.device, opt.pw_guide), opt)
# Optimizers
optimizer_G = torch.optim.Adam(generator.parameters(), lr=opt.lr, weight_decay=1e-8)
# Losses
# criterion_GAN = torch.nn.MSELoss()
# criterion_pixelwise = torch.nn.L1Loss()
# if opt.cuda:
# generator = generator.cuda()
# discriminator = discriminator.cuda()
generator, criterion_pixelwise = to_variables((generator,torch.nn.MSELoss()), cuda=opt.cuda, device=opt.device)
saverG = ModelSaver(f'{opt.checkpoint_dir}/saved_models/{opt.name}')
train_writer = SummaryWriter(log_dir=os.path.join(opt.checkpoint_dir, 'train'))
test_writer = SummaryWriter(log_dir=os.path.join(opt.checkpoint_dir, 'test'))
for epoch in tqdm(range(opt.epoch, opt.n_epochs), desc='Training'):
####
# Train
###
avg_stats = defaultdict(float)
for i, data in enumerate(train_loader):
data = to_variables(data, cuda=opt.cuda, device=opt.device)
y_hat, loss_G = trainG(generator, criterion_pixelwise, optimizer_G, data)
update_stats(avg_stats, loss_G)
# Print image to tensorboard
if (epoch % opt.sample_interval == 0) and (i % 50 == 0):
train_writer.add_image('RetouchNet', y_hat[0], epoch)
train_writer.add_image('Edited', data[2][0], epoch)
train_writer.add_image('Original', data[0][0], epoch)
# Log Progress
str_out = '[train] {}/{} '.format(epoch, opt.n_epochs)
for k, v in avg_stats.items():
avg = v / len(train_loader)
train_writer.add_scalar(k, avg, epoch)
str_out += '{}: {:.6f} '.format(k, avg)
print(str_out)
####
# Test
###
avg_stats = defaultdict(float)
images = None
with torch.no_grad():
for i, data in enumerate(test_loader):
data = to_variables(data, cuda=opt.cuda, device=opt.device, test=True)
images, losses = test(generator, criterion_pixelwise, data)
update_stats(avg_stats, losses)
# Print image to tensorboard
if (epoch % opt.sample_interval == 0) and (i % 5 == 0):
test_writer.add_image('RetouchNet', images[0], epoch)
test_writer.add_image('Edited', data[2][0], epoch)
test_writer.add_image('Original', data[0][0], epoch)
# Log Progress
str_out = '[test] {}/{} '.format(epoch, opt.n_epochs)
for k, v in avg_stats.items():
avg = v / len(test_loader)
test_writer.add_scalar(k, avg, epoch)
str_out += '{}: {:.6f} '.format(k, avg)
print(str_out)
# If at sample interval save image
# if epoch % opt.sample_interval == 0:
# x_hr, x_lr, y_hr, y_lr = data
# test_writer.add_image('RetouchNet', images[0], epoch)
# test_writer.add_image('GroundTruth', y_hr[0], epoch)
# test_writer.add_image('raw', x_hr[0], epoch)
if epoch % opt.checkpoint_interval == 0:
# Save model checkpoints
saverG.save_if_best(generator, loss_G['loss_G'])
def main():
db = vmdatabase.VMDatabase()
if not SKIP_DEVSTACK_GATE_JENKINS:
config = ConfigParser.ConfigParser()
config.read(DEVSTACK_GATE_SECURE_CONFIG)
jenkins = myjenkins.Jenkins(config.get('jenkins', 'server'),
config.get('jenkins', 'user'),
config.get('jenkins', 'apikey'))
jenkins.get_info()
else:
jenkins = None
print 'Known machines (start):'
db.print_state()
provider = db.getProvider(PROVIDER_NAME)
print "Working with provider %s" % provider.name
client = utils.get_client(provider)
flavor = utils.get_flavor(client, 1024)
print "Found flavor", flavor
error = False
now = time.time()
for machine in provider.machines:
# Normally, reap machines that have sat in their current state
# for 24 hours, unless that state is READY.
if (REAP_ALL_SERVERS
or (machine.state != vmdatabase.READY
and now - machine.state_time > MACHINE_LIFETIME)
or machine.state == vmdatabase.DELETE
or machine.state == vmdatabase.ERROR):
print 'Deleting machine', machine.name
try:
delete_machine(jenkins, client, machine)
except:
error = True
traceback.print_exc()
provider_min_ready = 0
for base_image in provider.base_images:
provider_min_ready += base_image.min_ready
for snap_image in base_image.snapshot_images:
# Normally, reap images that have sat in their current state
# for 24 hours, unless the image is the current snapshot
if (REAP_ALL_IMAGES
or (snap_image != base_image.current_snapshot
and now - snap_image.state_time > MACHINE_LIFETIME)):
print 'Deleting image', snap_image.name
try:
delete_image(client, snap_image)
except:
error = True
traceback.print_exc()
print
print 'Known machines (end):'
db.print_state()
utils.update_stats(provider)
if error:
sys.exit(1)
def forward(self,
images,
captions,
lengths,
img_lengths,
img_txts,
img_spans,
txt_spans,
labels,
ids=None,
epoch=None,
*args):
self.niter += 1
self.logger.update('Eit', self.niter)
self.logger.update('lr', self.optimizer.param_groups[0]['lr'])
img_lengths = torch.tensor(img_lengths).long() if isinstance(
img_lengths, list) else img_lengths
lengths = torch.tensor(lengths).long() if isinstance(lengths,
list) else lengths
if torch.cuda.is_available():
images = images.cuda()
captions = captions.cuda()
lengths = lengths.cuda()
img_lengths = img_lengths.cuda()
bsize = captions.size(0)
img_emb, nll_img, kl_img, span_margs_img, argmax_spans_img, trees_img, lprobs_img = self.forward_img_parser(
images, img_lengths)
ll_loss_img = nll_img.sum()
kl_loss_img = kl_img.sum()
txt_emb, nll_txt, kl_txt, span_margs_txt, argmax_spans_txt, trees_txt, lprobs_txt = self.forward_txt_parser(
captions, lengths)
ll_loss_txt = nll_txt.sum()
kl_loss_txt = kl_txt.sum()
contrastive_loss = self.forward_loss(img_emb, txt_emb, img_lengths,
lengths, argmax_spans_img,
argmax_spans_txt, span_margs_img,
span_margs_txt)
mt_loss = contrastive_loss.sum()
loss_img = self.vse_lm_alpha * (ll_loss_img + kl_loss_img) / bsize
loss_txt = self.vse_lm_alpha * (ll_loss_txt + kl_loss_txt) / bsize
loss_mt = self.vse_mt_alpha * mt_loss / bsize
loss = loss_img + loss_txt + loss_mt
self.optimizer.zero_grad()
loss.backward()
if self.grad_clip > 0:
clip_grad_norm_(self.all_params, self.grad_clip)
self.optimizer.step()
self.logger.update('Loss_img', loss_img.item(), bsize)
self.logger.update('Loss_txt', loss_txt.item(), bsize)
self.logger.update('KL-Loss_img', kl_loss_img.item() / bsize, bsize)
self.logger.update('KL-Loss_txt', kl_loss_txt.item() / bsize, bsize)
self.logger.update('LL-Loss_img', ll_loss_img.item() / bsize, bsize)
self.logger.update('LL-Loss_txt', ll_loss_txt.item() / bsize, bsize)
self.n_word_img += (img_lengths + 1).sum().item()
self.n_word_txt += (lengths + 1).sum().item()
self.n_sent += bsize
for b in range(bsize):
max_img_len = img_lengths[b].item()
pred_img = [(a[0], a[1]) for a in argmax_spans_img[b]
if a[0] != a[1]]
pred_set_img = set(pred_img[:-1])
gold_img = [(img_spans[b][i][0].item(), img_spans[b][i][1].item())
for i in range(max_img_len - 1)]
gold_set_img = set(gold_img[:-1])
utils.update_stats(pred_set_img, [gold_set_img],
self.all_stats_img)
max_txt_len = lengths[b].item()
pred_txt = [(a[0], a[1]) for a in argmax_spans_txt[b]
if a[0] != a[1]]
pred_set_txt = set(pred_txt[:-1])
gold_txt = [(txt_spans[b][i][0].item(), txt_spans[b][i][1].item())
for i in range(max_txt_len - 1)]
gold_set_txt = set(gold_txt[:-1])
utils.update_stats(pred_set_txt, [gold_set_txt],
self.all_stats_txt)
# if self.niter % self.log_step == 0:
p_norm, g_norm = self.norms()
all_f1_img = utils.get_f1(self.all_stats_img)
all_f1_txt = utils.get_f1(self.all_stats_txt)
train_kl_img = self.logger.meters["KL-Loss_img"].sum
train_ll_img = self.logger.meters["LL-Loss_img"].sum
train_kl_txt = self.logger.meters["KL-Loss_txt"].sum
train_ll_txt = self.logger.meters["LL-Loss_txt"].sum
info = '|Pnorm|: {:.6f}, |Gnorm|: {:.2f}, ReconPPL-Img: {:.2f}, KL-Img: {:.2f}, ' + \
'PPLBound-Img: {:.2f}, CorpusF1-Img: {:.2f}, ' + \
'ReconPPL-Txt: {:.2f}, KL-Txt: {:.2f}, ' + \
'PPLBound-Txt: {:.2f}, CorpusF1-Txt: {:.2f}, ' + \
'Speed: {:.2f} sents/sec'
info = info.format(
p_norm, g_norm, np.exp(train_ll_img / self.n_word_img),
train_kl_img / self.n_sent,
np.exp((train_ll_img + train_kl_img) / self.n_word_img),
all_f1_img[0], np.exp(train_ll_txt / self.n_word_txt),
train_kl_txt / self.n_sent,
np.exp((train_ll_txt + train_kl_txt) / self.n_word_txt),
all_f1_txt[0], self.n_sent / (time.time() - self.s_time))
pred_action_img = utils.get_actions(trees_img[0])
sent_s_img = img_txts[0]
pred_t_img = utils.get_tree(pred_action_img, sent_s_img)
gold_t_img = utils.span_to_tree(img_spans[0].tolist(),
img_lengths[0].item())
gold_action_img = utils.get_actions(gold_t_img)
gold_t_img = utils.get_tree(gold_action_img, sent_s_img)
info += "\nPred T Image: {}\nGold T Image: {}".format(
pred_t_img, gold_t_img)
pred_action_txt = utils.get_actions(trees_txt[0])
sent_s_txt = [
self.vocab.idx2word[wid] for wid in captions[0].cpu().tolist()
]
pred_t_txt = utils.get_tree(pred_action_txt, sent_s_txt)
gold_t_txt = utils.span_to_tree(txt_spans[0].tolist(),
lengths[0].item())
gold_action_txt = utils.get_actions(gold_t_txt)
gold_t_txt = utils.get_tree(gold_action_txt, sent_s_txt)
info += "\nPred T Text: {}\nGold T Text: {}".format(
pred_t_txt, gold_t_txt)
return info