parser.add_argument("--model_file", required=True, help='model file to use (e.g. model_best.pt)')
parser.add_argument("--bert_model_file", required=True, help='bert model file to use (e.g. model_bert_best.pt)')
parser.add_argument("--bert_path", required=True, help='path to bert files (bert_config*.json etc)')
parser.add_argument("--data_path", required=True, help='path to *.jsonl and *.db files')
parser.add_argument("--split", required=True, help='prefix of jsonl and db files (e.g. dev)')
parser.add_argument("--result_path", required=True, help='directory in which to place results')
args = construct_hyper_param(parser)
BERT_PT_PATH = args.bert_path
path_save_for_evaluation = args.result_path
# Load pre-trained models
path_model_bert = args.bert_model_file
path_model = args.model_file
args.no_pretraining = True # counterintuitive, but avoids loading unused models
model, model_bert, tokenizer, bert_config = get_models(args, BERT_PT_PATH, trained=True, path_model_bert=path_model_bert, path_model=path_model)
# Load data
dev_data, dev_table = load_wikisql_data(args.data_path, mode=args.split, toy_model=False, toy_size=args.toy_size, no_hs_tok=True)
dev_loader = torch.utils.data.DataLoader(
batch_size=args.bS,
dataset=dev_data,
shuffle=False,
num_workers=1,
collate_fn=lambda x: x # now dictionary values are not merged!
)
if(args.split=='test'):
dev_data.remove(dev_data[1884])
dev_data.remove(dev_data[1883])
dev_data.remove(dev_data[1882])
if not os.path.exists(path_save_for_evaluation) :
# please provide the path to your training and testing datasets
training = pd.read_csv("...")
test = pd.read_csv("....")
complications = ['SBI', 'AKI', 'ARDS']
framework_train = apply_stratified_framework(training, complications)
framework_test = apply_stratified_framework(test, complications)
train_columns = [
'Diastolic Blood Pressure_mean', 'Diastolic Blood Pressure_min',
'Oxygen Saturation_max', 'Oxygen Saturation_mean', 'Oxygen Saturation_min',
'Peripheral Pulse Rate_max', 'Peripheral Pulse Rate_mean',
'Peripheral Pulse Rate_min', 'Respiratory Rate_max',
'Respiratory Rate_mean', 'Respiratory Rate_min',
'Systolic Blood Pressure_max', 'Systolic Blood Pressure_mean',
'Systolic Blood Pressure_min', 'Temperature Axillary_max',
'Temperature Axillary_mean', 'Temperature Axillary_min', 'GCS_mean',
'GCS_min', 'GCS_max', 'GENDER', 'AGE', 'COUGH', 'FEVER', 'SOB',
'SORE_THROAT', 'RASH', 'BMI', 'DIABETES', 'HYPERTENSION', 'CKD', 'CANCER'
]
models_all, trainsets, classifers = get_models(complications, framework_train,
train_columns)
true_ouctomes, predicted_ouctomes = get_results(framework_test, complications,
models_all, train_columns)
plot_roc(complications, true_ouctomes, predicted_ouctomes, "testset")
plot_PRC(complications, true_ouctomes, predicted_ouctomes, "testset")
temporal_transform=temporal_transform['train'],
openpose_transform=openpose_transform['train'],
spatial_transform=spatial_transform['train'],
arguments=args),
'val':
HandHygiene(os.path.join(VIDEO_DIR, 'val'),
temporal_transform=temporal_transform['val'],
openpose_transform=openpose_transform['val'],
spatial_transform=spatial_transform['val'],
arguments=args),
}
# create model
i3d_rgb, i3d_flow = get_models(len(args.label),
True,
170,
load_pt_weights=True,
rgb_weights_path=args.model_path.rgb,
flow_weights_path=args.model_path.flow)
if torch.cuda.device_count() > 1:
i3d_rgb = torch.nn.DataParallel(i3d_rgb).cuda()
i3d_flow = torch.nn.DataParallel(i3d_flow).cuda()
# hyperparameters / trainable parameters
optims = {'rgb': None, 'flow': None}
schedulers = {'rgb': None, 'flow': None}
feature_extract = True
def trainable_params(model, mode='rgb'):
params_to_update = model.parameters()
print("Params to learn:")