def launch():
global current_render_distance, current_username, current_process
launch_button.state(['disabled'])
if current_process is None or current_process.poll() is not None:
current_process = launcher.run(get_features(),
current_render_distance.get(),
current_username.get())
return 0
def sequentialStep():
print('Starting with ', globalvars.appName)
# creates child processes, each is a sniper run. sim.out is saved in
# individual directories
print('++Launcher++')
launcher.run(permutations, parameterNames, False, True)
# iterates over the sim.out and macpat files in the relevant directories to
# collect stats. Stats are logged in a file 'efficiency.txt' in app directory
print('++Stat-Collection++')
stats.collectStats(permutations)
# now that we have data for all the configurations, we can select the 'good' ones
print('++Selection++')
goodConfigs = selection.pickGood([x[1:] for x in parameters])
goodConfigDict[globalvars.appName] = goodConfigs
def parallelDetailed(appName, pinballLoc, outputDirBase):
# log the output of the process to a specific file
sys.stdout = open(outputDirBase+"plog.out", "w")
appfeatures = list(list())
#append the names of the features
appfeatures.append(featureNames)
goodConfigs = goodConfigDict[appName]
# now that we have the good configs, we need to launch sniper runs again for
# these configs with detailed stat (counter-data) collection enabled
#print('++Detailed-Run++')
launcher.run(goodConfigs, parameterNames, True, False, pinballLoc, outputDirBase)
# from the detailed runs, we now extract the feature vectors - this forms the
# training data. Feature vectors are logged to a file
#print('++Extraction++')
featureGeneration.extract(goodConfigs, appfeatures, outputDirBase)
# Launcher
# http://blog.tvalacarta.info/plugin-xbmc/tvalacarta/
#------------------------------------------------------------
# Constantes
__plugin__ = "tvalacarta"
__author__ = "tvalacarta"
__url__ = "http://blog.tvalacarta.info/plugin-xbmc/tvalacarta/"
__date__ = "1 Septiembre 2010"
__version__ = "2.6"
import os
import sys
import xbmc
xbmc.output("[default.py] tvalacarta init...")
# Configura los directorios donde hay librerías
#librerias = xbmc.translatePath( os.path.join( os.getcwd(), 'resources', 'lib' ) )
#sys.path.append (librerias)
#librerias = xbmc.translatePath( os.path.join( os.getcwd(), 'channels' ) )
#sys.path.append (librerias)
#librerias = xbmc.translatePath( os.path.join( os.getcwd(), 'servers' ) )
#sys.path.append (librerias)
librerias = xbmc.translatePath( os.path.join( os.getcwd(), 'youtubeAPI' ) )
sys.path.append (librerias)
# Ejecuta el programa principal
import launcher
launcher.run()
f.detach().cpu()],
dim=3)
canvas.append(r)
return canvas
canvas = make(paths[:40])
canvas = torch.cat(canvas, dim=0)
save_image(canvas * 0.5 + 0.5,
'make_figures/output/reconstructions_bed_1.png',
nrow=4,
pad_value=1.0)
canvas = make(paths[40:80])
canvas = torch.cat(canvas, dim=0)
save_image(canvas * 0.5 + 0.5,
'make_figures/output/reconstructions_bed_2.png',
nrow=4,
pad_value=1.0)
if __name__ == "__main__":
gpu_count = 1
run(sample,
get_cfg_defaults(),
description='ALAE-reconstruction-bedroom',
default_config='configs/bedroom.yaml',
world_size=gpu_count,
write_log=False)
langevin_steps=cfg.LANGEVIN.STEP,
lr=cfg.LANGEVIN.LR)
with torch.no_grad():
latents = torch.cat((source_latent[:nrow], latents))
latents = latents.unsqueeze(1).repeat(1,
ae.mapping_fl.num_layers,
1)
out = decode(ae, latents, cfg)
out = torch.cat((source_img[:nrow], out), dim=0)
utils.save_image(
out,
f"{run_dir}/{str(iterations).zfill(6)}.png",
nrow=nrow,
normalize=True,
padding=0,
range=(-1, 1),
)
if __name__ == "__main__":
gpu_count = 1
run(train,
get_cfg_defaults(),
description='Image-Translation',
default_config='configs/celeba-hq.yaml',
world_size=gpu_count,
write_log=False)
model.eval()
layer_count = cfg.MODEL.LAYER_COUNT
decoder = nn.DataParallel(decoder)
im_size = 2**(cfg.MODEL.LAYER_COUNT + 1)
with torch.no_grad():
draw_uncurated_result_figure(cfg,
'make_figures/output/%s/generations.jpg' %
cfg.NAME,
model,
cx=0,
cy=0,
cw=im_size,
ch=im_size,
rows=6,
lods=[0, 0, 0, 1, 1, 2],
seed=5)
if __name__ == "__main__":
gpu_count = 1
run(sample,
get_cfg_defaults(),
description='ALAE-generations',
default_config='configs/ffhq.yaml',
world_size=gpu_count,
write_log=False)
'dlatent_avg': dlatent_avg
}
checkpointer = Checkpointer(cfg, model_dict, {}, logger=logger, save=False)
checkpointer.load()
model.eval()
layer_count = cfg.MODEL.LAYER_COUNT
logger.info("Generating...")
decoder = nn.DataParallel(decoder)
mapping_fl = nn.DataParallel(mapping_fl)
with torch.no_grad():
gen = ImageGenerator(cfg, num_samples=60000, minibatch_gpu=8)
gen.evaluate(logger, mapping_fl, decoder,
cfg.DATASET.MAX_RESOLUTION_LEVEL - 2)
if __name__ == "__main__":
gpu_count = 1
run(sample,
get_cfg_defaults(),
description='ALAE-generate-images-for-attribute-classifications',
default_config='configs/ffhq.yaml',
world_size=gpu_count,
write_log=False)
extra_checkpoint_data = checkpointer.load()
last_epoch = list(extra_checkpoint_data['auxiliary']
['scheduler'].values())[0]['last_epoch']
logger.info("Model trained for %d epochs" % last_epoch)
model.eval()
layer_count = cfg.MODEL.LAYER_COUNT
logger.info("Evaluating FID metric")
model.decoder = nn.DataParallel(decoder)
with torch.no_grad():
ppl = FID(cfg,
num_images=50000,
minibatch_size=16 * torch.cuda.device_count())
ppl.evaluate(logger, mapping_fl, model.decoder, model,
cfg.DATASET.MAX_RESOLUTION_LEVEL - 2)
if __name__ == "__main__":
gpu_count = 1
run(sample,
get_cfg_defaults(),
description='ALAE-fid',
default_config='configs/ffhq.yaml',
world_size=gpu_count,
write_log="metrics/fid_score.txt")
def launchEmulator(self): path=r'\Symbian\9.1\S60_3rd\Epoc32\release\winscw\udeb\bluewhaleplatform.exe' print "launch ", path child = launcher.run(path) return child
def launchEmulator(self):
path = r'\Symbian\9.1\S60_3rd\Epoc32\release\winscw\udeb\bluewhaleplatform.exe'
print "launch ", path
child = launcher.run(path)
return child
model.eval()
im_size = 2**(cfg.MODEL.LAYER_COUNT + 1)
seed = np.random.randint(0, 999999)
print("seed:", seed)
with torch.no_grad():
path = './make_figures/output'
os.makedirs(path, exist_ok=True)
os.makedirs(os.path.join(path, cfg.NAME), exist_ok=True)
draw_uncurated_result_figure(
cfg,
'./make_figures/output/%s/generations.jpg' % cfg.NAME,
model,
cx=0,
cy=0,
cw=im_size,
ch=im_size,
rows=6,
lods=[0, 0, 0, 1, 1, 2],
seed=seed)
if __name__ == "__main__":
gpu_count = 1
run(sample,
get_cfg_defaults(),
description='SoftIntroVAE-generations',
default_config='./configs/ffhq256.yaml',
world_size=gpu_count,
write_log=False)
requires_grad=True).cuda() / 127.5 - 1.
if x.shape[0] == 4:
x = x[:3]
while x.shape[2] != model.decoder.layer_to_resolution[6]:
x = F.avg_pool2d(x, 2, 2)
latents = encode(x[None, ...].cuda())
f = decode(latents)
r = torch.cat([x[None, ...].detach().cpu(),
f.detach().cpu()],
dim=3)
os.makedirs('make_figures/output/pioneer/', exist_ok=True)
save_image(f.detach().cpu() * 0.5 + 0.5,
'make_figures/output/pioneer/%s_alae.png' %
filename[:-9],
nrow=1,
pad_value=1.0)
make(paths)
if __name__ == "__main__":
gpu_count = 1
run(sample,
get_cfg_defaults(),
description='ALAE-reconstructions-celeb-hq256-on-pioneer-examples',
default_config='configs/celeba-hq256.yaml',
world_size=gpu_count,
write_log=False)
lod2batch.step()
# if lod2batch.is_time_to_save():
# checkpointer.save("model_tmp_intermediate_lod%d" % lod_for_saving_model)
if lod2batch.is_time_to_report():
save_sample(lod2batch, tracker, sample, samplez, x, logger, model_s, cfg, encoder_optimizer,
generator_optimizer, output_folder)
scheduler.step()
if epoch % 20 == 0:
save(epoch)
save_sample(lod2batch, tracker, sample, samplez, x, logger, model_s, cfg, encoder_optimizer, generator_optimizer, output_folder)
logger.info("Training finish!... save training results")
if epoch is not None:
save(epoch)
best_model_name, best_model_score = scores_list[0]
for model_name, model_score in scores_list:
if model_score >= best_model_score:
best_model_name, best_model_score = model_name, model_score
checkpointer.tag_best_checkpoint(best_model_name)
if __name__ == "__main__":
gpu_count = torch.cuda.device_count()
run(train, get_cfg_defaults(), description='', default_config='configs/mnist_os.yaml',
world_size=gpu_count)
lod_for_saving_model = lod2batch.lod
lod2batch.step()
if local_rank == 0:
if lod2batch.is_time_to_save():
checkpointer.save("model_tmp_intermediate_lod%d" %
lod_for_saving_model)
if lod2batch.is_time_to_report():
save_sample(lod2batch, tracker, sample, samplez, x, logger,
model_s, cfg, encoder_optimizer,
decoder_optimizer)
scheduler.step()
if local_rank == 0:
checkpointer.save("model_tmp_lod%d" % lod_for_saving_model)
save_sample(lod2batch, tracker, sample, samplez, x, logger,
model_s, cfg, encoder_optimizer, decoder_optimizer)
logger.info("Training finish!... save training results")
if local_rank == 0:
checkpointer.save("model_final").wait()
if __name__ == "__main__":
gpu_count = torch.cuda.device_count()
run(train,
get_cfg_defaults(),
description='StyleGAN',
default_config='configs/ffhq.yaml',
world_size=gpu_count)
for i in range(src_len):
save_image(src_originals[i] * 0.5 + 0.5, 'style_mixing/output/%s/source_%d.png' % (cfg.NAME, i))
place(canvas, src_originals[i], 1 + i, 0)
for i in range(dst_len):
save_image(dst_originals[i] * 0.5 + 0.5, 'style_mixing/output/%s/dst_coarse_%d.png' % (cfg.NAME, i))
place(canvas, dst_originals[i], 0, 1 + i)
style_ranges = [range(0, 4)] * 3 + [range(4, 8)] * 2 + [range(8, layer_count * 2)]
def mix_styles(style_src, style_dst, r):
style = style_dst.clone()
style[:, r] = style_src[:, r]
return style
for row in range(dst_len):
row_latents = torch.stack([dst_latents[row]] * src_len)
style = mix_styles(src_latents, row_latents, style_ranges[row])
rec = model.decoder(style, layer_count - 1, 1, noise=True)
for j in range(rec.shape[0]):
save_image(rec[j] * 0.5 + 0.5, 'style_mixing/output/%s/rec_coarse_%d_%d.png' % (cfg.NAME, row, j))
place(canvas, rec[j], 1 + j, 1 + row)
save_image(torch.Tensor(canvas), 'style_mixing/output/%s/stylemix.png' % cfg.NAME)
if __name__ == "__main__":
gpu_count = 1
run(main, get_cfg_defaults(), description='ALAE-style-mixing', default_config='configs/ffhq.yaml',
world_size=gpu_count, write_log=False)
# Launcher
# http://blog.tvalacarta.info/plugin-xbmc/tvalacarta/
#------------------------------------------------------------
# Constantes
__plugin__ = "tvalacarta"
__author__ = "tvalacarta"
__url__ = "http://blog.tvalacarta.info/plugin-xbmc/tvalacarta/"
__date__ = "1 Septiembre 2010"
__version__ = "2.6"
import os
import sys
import xbmc
xbmc.output("[default.py] tvalacarta init...")
# Configura los directorios donde hay librerías
#librerias = xbmc.translatePath( os.path.join( os.getcwd(), 'resources', 'lib' ) )
#sys.path.append (librerias)
#librerias = xbmc.translatePath( os.path.join( os.getcwd(), 'channels' ) )
#sys.path.append (librerias)
#librerias = xbmc.translatePath( os.path.join( os.getcwd(), 'servers' ) )
#sys.path.append (librerias)
librerias = xbmc.translatePath(os.path.join(os.getcwd(), 'youtubeAPI'))
sys.path.append(librerias)
# Ejecuta el programa principal
import launcher
launcher.run()
extra_checkpoint_data = checkpointer.load()
last_epoch = list(extra_checkpoint_data['auxiliary']
['scheduler'].values())[0]['last_epoch']
logger.info("Model trained for %d epochs" % last_epoch)
model.eval()
layer_count = cfg.MODEL.LAYER_COUNT
logger.info("Evaluating LPIPS metric")
decoder = nn.DataParallel(decoder)
encoder = nn.DataParallel(encoder)
with torch.no_grad():
ppl = LPIPS(cfg,
num_images=10000,
minibatch_size=16 * torch.cuda.device_count())
ppl.evaluate(logger, mapping_fl, decoder, encoder,
cfg.DATASET.MAX_RESOLUTION_LEVEL - 2)
if __name__ == "__main__":
gpu_count = 1
run(sample,
get_cfg_defaults(),
description='ALAE-lpips',
default_config='configs/experiment_celeba.yaml',
world_size=gpu_count,
write_log="metrics/lpips_score.txt")
novelty_detector = model_s, bin_edges, counts, gennorm_param,
percentages = cfg.DATASET.PERCENTAGES
# percentages = [50]
results = {}
for p in percentages:
# plt.figure(num=None, figsize=(8, 6), dpi=180, facecolor='w', edgecolor='k')
alpha, beta, threshold, _ = compute_threshold_coeffs(
cfg, logger, valid_set, inliner_classes, p, novelty_detector)
with open(
os.path.join(output_folder,
'coeffs_percentage_%d.txt' % int(p)), 'w') as f:
f.write("%f %f %f\n" % (alpha, beta, threshold))
results[p] = test(cfg, logger, test_set, inliner_classes, p,
novelty_detector, alpha, beta, threshold,
output_folder)
return results
if __name__ == "__main__":
run(main,
get_cfg_defaults(),
description='',
default_config='configs/mnist.yaml',
world_size=1,
folding_id=0,
inliner_classes=[3])
canvas.append(r)
return canvas
def chunker_list(seq, n):
return [seq[i * n:(i + 1) * n] for i in range((len(seq) + n - 1) // n)]
paths = chunker_list(paths, 8 * 3)
path = './make_figures/output'
os.makedirs(path, exist_ok=True)
os.makedirs(os.path.join(path, cfg.NAME), exist_ok=True)
for i, chunk in enumerate(paths):
canvas = make(chunk)
canvas = torch.cat(canvas, dim=0)
save_path = './make_figures/output/%s/reconstructions_%d.png' % (
cfg.NAME, i)
os.makedirs(os.path.dirname(save_path), exist_ok=True)
save_image(canvas * 0.5 + 0.5, save_path, nrow=3, pad_value=1.0)
if __name__ == "__main__":
gpu_count = 1
run(sample,
get_cfg_defaults(),
description='SoftIntroVAE-figure-reconstructions-paged',
default_config='./configs/ffhq256.yaml',
world_size=gpu_count,
write_log=False)
save_image(src_originals[i] * 0.5 + 0.5, './style_mixing/output/%s/source_%d.png' % (cfg.NAME, i))
place(canvas, src_originals[i], 1 + i, 0)
for i in range(dst_len):
save_image(dst_originals[i] * 0.5 + 0.5, './style_mixing/output/%s/dst_coarse_%d.png' % (cfg.NAME, i))
place(canvas, dst_originals[i], 0, 1 + i)
style_ranges = [range(0, 4)] * 3 + [range(4, 8)] * 2 + [range(8, layer_count * 2)]
def mix_styles(style_src, style_dst, r):
style = style_dst.clone()
style[:, r] = style_src[:, r]
return style
for row in range(dst_len):
row_latents = torch.stack([dst_latents[row]] * src_len)
style = mix_styles(src_latents, row_latents, style_ranges[row])
rec = model.decoder(style, layer_count - 1, 1, noise=True)
for j in range(rec.shape[0]):
save_image(rec[j] * 0.5 + 0.5, './style_mixing/output/%s/rec_coarse_%d_%d.png' % (cfg.NAME, row, j))
place(canvas, rec[j], 1 + j, 1 + row)
save_image(torch.Tensor(canvas), './style_mixing/output/%s/stylemix.png' % cfg.NAME)
if __name__ == "__main__":
gpu_count = 1
run(main, get_cfg_defaults(), description='SandwichStyleVAE-style-mixing',
default_config='./configs/celeba-hq256-generate.yaml',
world_size=gpu_count, write_log=False)
lod_for_saving_model = lod2batch.lod
lod2batch.step()
if local_rank == 0:
if lod2batch.is_time_to_save():
checkpointer.save("model_tmp_intermediate_lod%d" %
lod_for_saving_model)
if lod2batch.is_time_to_report():
save_sample(lod2batch, tracker, sample, samplez, x,
logger, model_s, cfg, encoder_optimizer,
decoder_optimizer)
scheduler.step()
if local_rank == 0:
checkpointer.save("model_tmp_lod%d" % lod_for_saving_model)
save_sample(lod2batch, tracker, sample, samplez, x, logger,
model_s, cfg, encoder_optimizer, decoder_optimizer)
logger.info("Training finish!... save training results")
if local_rank == 0:
checkpointer.save("model_final").wait()
if __name__ == "__main__":
gpu_count = torch.cuda.device_count()
run(train,
get_cfg_defaults(),
description='StyleGAN',
default_config='configs/experiment_celeba_sep.yaml',
world_size=gpu_count)
save_sample(lod2batch, tracker, sample, samplez, x, logger,
model_s, cfg, encoder_optimizer,
decoder_optimizer)
scheduler.step()
mean_diff_kl = np.mean(diff_kls)
print("mean diff kl: ", mean_diff_kl)
if epoch > num_vae_epochs - 1:
kls_real.append(np.mean(batch_kls_real))
kls_fake.append(np.mean(batch_kls_fake))
rec_errs.append(np.mean(batch_rec_errs))
if local_rank == 0:
checkpointer.save("model_tmp_lod%d" % lod_for_saving_model)
save_sample(lod2batch, tracker, sample, samplez, x, logger,
model_s, cfg, encoder_optimizer, decoder_optimizer)
logger.info("Training finish!... save training results")
if local_rank == 0:
checkpointer.save("model_final").wait()
if __name__ == "__main__":
gpu_count = torch.cuda.device_count()
run(train,
get_cfg_defaults(),
description='StyleSoftIntroVAE',
default_config='./configs/ffhq256.yaml',
world_size=gpu_count)