def hello_world(args):
"""
Runs Jina's Hello World.
Usage:
Use it via CLI :command:`jina hello-world`.
Description:
It downloads Fashion-MNIST dataset and :term:`Indexer<indexes>` 50,000 images.
The index is stored into 4 *shards*. It randomly samples 128 unseen images as :term:`Queries<Searching>`
Results are shown in a webpage.
More options can be found in :command:`jina hello-world --help`
:param args: Argparse object
"""
Path(args.workdir).mkdir(parents=True, exist_ok=True)
targets = {
'index-labels': {
'url': args.index_labels_url,
'filename': os.path.join(args.workdir, 'index-labels'),
},
'query-labels': {
'url': args.query_labels_url,
'filename': os.path.join(args.workdir, 'query-labels'),
},
'index': {
'url': args.index_data_url,
'filename': os.path.join(args.workdir, 'index-original'),
},
'query': {
'url': args.query_data_url,
'filename': os.path.join(args.workdir, 'query-original'),
},
}
# download the data
download_data(targets, args.download_proxy)
# reduce the network load by using `fp16`, or even `uint8`
os.environ['JINA_ARRAY_QUANT'] = 'fp16'
os.environ['HW_WORKDIR'] = args.workdir
# now comes the real work
# load index flow from a YAML file
f = Flow().add(uses=MyEncoder,
parallel=2).add(uses=MyIndexer).add(uses=MyEvaluator)
# run it!
with f:
f.index(
index_generator(num_docs=targets['index']['data'].shape[0],
target=targets),
request_size=args.request_size,
)
# wait for couple of seconds
countdown(
3,
reason=colored(
'behold! im going to switch to query mode',
'cyan',
attrs=['underline', 'bold', 'reverse'],
),
)
# f.search(
# query_generator(
# num_docs=args.num_query, target=targets, with_groundtruth=True
# ),
# shuffle=True,
# on_done=print_result,
# request_size=args.request_size,
# parameters={'top_k': args.top_k},
# )
f.post(
'/eval',
query_generator(num_docs=args.num_query,
target=targets,
with_groundtruth=True),
shuffle=True,
on_done=print_result,
request_size=args.request_size,
parameters={'top_k': args.top_k},
)
# write result to html
write_html(os.path.join(args.workdir, 'demo.html'))
def hello_world(args):
"""
Runs Jina's Hello World.
Usage:
Use it via CLI :command:`jina hello fashion`.
Description:
It downloads Fashion-MNIST dataset and :term:`Indexer<indexes>` 50,000 images.
The index is stored into 4 *shards*. It randomly samples 128 unseen images as :term:`Queries<Searching>`
Results are shown in a webpage.
More options can be found in :command:`jina hello-world --help`
:param args: Argparse object
"""
Path(args.workdir).mkdir(parents=True, exist_ok=True)
targets = {
'index-labels': {
'url': args.index_labels_url,
'filename': os.path.join(args.workdir, 'index-labels'),
},
'query-labels': {
'url': args.query_labels_url,
'filename': os.path.join(args.workdir, 'query-labels'),
},
'index': {
'url': args.index_data_url,
'filename': os.path.join(args.workdir, 'index-original'),
},
'query': {
'url': args.query_data_url,
'filename': os.path.join(args.workdir, 'query-original'),
},
}
# download the data
download_data(targets, args.download_proxy)
# now comes the real work
# load index flow from a YAML file
f = (Flow().add(uses=MyEncoder, replicas=2).add(uses=MyIndexer,
workspace=args.workdir))
# run it!
with f:
f.index(
index_generator(num_docs=targets['index']['data'].shape[0],
target=targets),
show_progress=True,
)
groundtruths = get_groundtruths(targets)
evaluate_print_callback = partial(print_result, groundtruths)
evaluate_print_callback.__name__ = 'evaluate_print_callback'
f.post(
'/search',
query_generator(num_docs=args.num_query, target=targets),
shuffle=True,
on_done=evaluate_print_callback,
parameters={'top_k': args.top_k},
show_progress=True,
)
# write result to html
write_html(os.path.join(args.workdir, 'demo.html'))
if not os.path.exists(log_dir): os.mkdir(log_dir)
FLAGS.model = model
FLAGS.data_dict = data_dict
FLAGS.model_path = model_path
FLAGS.log_dir = log_dir
FLAGS.pairs_file = pairs_file
FLAGS.eval_poses = eval_poses
set_p = store_params(FLAGS)
return set_p
if __name__ == '__main__':
if set_params(FLAGS.model, FLAGS.data_dict, FLAGS.model_path,
FLAGS.log_dir, FLAGS.pairs_file, FLAGS.eval_poses):
MODEL = importlib.import_module(FLAGS.model) # import network module
helper.download_data(FLAGS.data_dict)
train()
# Used the following code to test multiple datas and multiple logs.
# model_paths2test = ['log_multi_catg/model250.ckpt']#,['log_multi_catg_noise/model300.ckpt']
# # model_paths2test = ['log_car_multi_models/model200.ckpt']#, 'log_car_multi_models_noise/model350.ckpt']
# data_dicts2test = ['train_data','unseen_data']#, 'car_data']#, 'unseen_data']
# eval_poses2test = ['itr_net_test_data45.csv']
# FLAGS.use_noise_data = False
# for mpt in model_paths2test:
# for ddt in data_dicts2test:
# for ept in eval_poses2test:
# log_dir = 'test_itr_'+str(mpt[4:len(mpt)-14])+'_'+str(ddt)+'_'+str(ept[len(ept)-6:len(ept)-4])
eval_writer.add_summary(summary, step) # Add all the summary to the tensorboard.
# Apply the final transformation on the template data and multiply it with the transformation matrix obtained from N-Iterations.
TRANSFORMATIONS, source_data = helper.transformation_quat2mat(predicted_transformation, TRANSFORMATIONS, source_data)
final_pose = helper.find_final_pose_inv(TRANSFORMATIONS) # Find the final pose (translation, orientation (euler angles in degrees)) from transformation matrix.
# Display the ground truth pose and predicted pose for first Point Cloud in batch
if display_poses:
print('Ground Truth Position: {}'.format(batch_euler_poses[0,0:3].tolist()))
print('Predicted Position: {}'.format(final_pose[0,0:3].tolist()))
print('Ground Truth Orientation: {}'.format((batch_euler_poses[0,3:6]*(180/np.pi)).tolist()))
print('Predicted Orientation: {}'.format((final_pose[0,3:6]*(180/np.pi)).tolist()))
# Display Loss Value.
print("Batch: {}, Loss: {}\r".format(fn, loss_val),end='')
# Add loss for each batch.
loss_sum += loss_val
print('\n')
log_string('Eval Mean loss: %f' % (loss_sum/num_batches)) # Store and display mean loss of epoch.
if __name__ == "__main__":
if FLAGS.mode == 'no_mode':
print('Specity a mode argument: train')
elif FLAGS.mode == 'train':
if helper.download_data(FLAGS.data_dict): print_('################### Data Downloading Finished ###################', color='g', style='bold')
train()
LOG_FOUT.close()
