def main(args):
## set the gpu
if args.gpu:
os.environ["CUDA_VISIBLE_DEVICES"] = args.gpu
## check the ckpt path
path_create('ckpt')
## check the model path
path_create('models')
## check the data path
path_create('data')
## check the log path
path_create('log')
if args.type == 'train':
train_main(args)
elif args.type == 'eval':
eval_main(args)
def voc_test(eval_steps):
train_main('latest.pth', split='trainaug')
eval_main('best_trainaug.pth', eval_steps, 'best')
eval_main('final_trainaug.pth', eval_steps, 'final')
settings.LR = settings.LR / 10
settings.ITER_MAX = settings.TEST_ITER_MAX
settings.ITER_SAVE = settings.TEST_ITER_SAVE
settings.ITER_VAL = settings.TEST_ITER_VAL
train_main('final_trainaug.pth',
split='trainval',
val=True,
reset_steps=True)
eval_steps = [5, 6, 7]
eval_main('best.pth', eval_steps, 'trainval_best')
eval_main('final.pth', eval_steps, 'trainval_final')
parser.add_argument("--log", type=int, default=log)
# float
parser.add_argument("--lr", type=float, default=lr)
parser.add_argument("--beta", type=float, default=beta)
parser.add_argument("--dropout_prob", type=float, default=dropout_prob)
args = parser.parse_args()
# overwrite params
midi_op = args.midi_op
process_collate = args.process_collate
input_type = args.input_type
band = args.band
split = args.split
score_choose = args.score_choose
batch_size = args.batch_size
sample_num = args.sample_num
chunk_size = args.chunk_size
lr = args.lr
log = bool(args.log)
beta = args.beta
num_rec_layers = args.num_rec_layers
z_dim = args.z_dim
kernel_size = args.kernel_size
stride = args.stride
num_conv_features = args.num_conv_features
train_main()
eval_main()
#Run eval
#Always use all hand annotations
experiment = Experiment(api_key="ypQZhYfs3nSyKzOfz13iuJpj2",
project_name='deeplidar',
log_code=False)
experiment.log_parameter("mode", "evaluation_grid")
experiment.log_parameters(DeepForest_config)
args = [
"--batch-size",
str(DeepForest_config['batch_size']), '--score-threshold',
str(DeepForest_config['score_threshold']),
'--suppression-threshold', '0.1', '--save-path',
'snapshots/images/', '--model', model, '--convert-model'
]
stem_recall, mAP = eval_main(data=data,
DeepForest_config=DeepForest_config,
experiment=experiment,
args=args)
results.append({
"Evaluation Site": site,
"Pretraining Site": pretraining_site,
"Stem Recall": stem_recall,
"mAP": mAP
})
results = pd.DataFrame(results)
results.to_csv("analysis/site_grid" + ".csv")
def voc_val(eval_steps):
train_main('latest.pth', split='trainaug')
eval_main('best_trainaug.pth', eval_steps, 'best')
eval_main('final_trainaug.pth', eval_steps, 'final')
data,
DeepForest_config,
experiment=experiment)
#Format output
experiment = Experiment(api_key="ypQZhYfs3nSyKzOfz13iuJpj2",
project_name='deeplidar',
log_code=False)
experiment.log_parameter("mode", "retrain_sequence_evaluation")
#pass an args object instead of using command line
retinanet_args = [
"--batch-size",
str(DeepForest_config['batch_size']), '--score-threshold',
str(DeepForest_config['score_threshold']), '--suppression-threshold',
'0.1', '--save-path', 'snapshots/images/', '--model', trained_model,
'--convert-model'
]
mAP = eval_main(data=data,
DeepForest_config=DeepForest_config,
experiment=experiment,
args=retinanet_args)
model_name = os.path.splitext(os.path.basename(model))[0]
results.append({"Model": model_name, "mAP": mAP})
results = pd.DataFrame(results)
print(results)
results.to_csv("pretraining_size" + ".csv")