from pathlib import Path
from utils_model import get_predictions
# Run the ResNet on the generated patches.
print("\n\n+++++ Running 4_test.py +++++")
print("\n----- Finding validation patch predictions -----")
# Validation patches.
get_predictions(patches_eval_folder=Path(
"/home/ifsdata/vlg/jason/easy-self-supervised-training/data/voc_trainval_full/val"
),
output_folder=Path("outputs").joinpath(
"resnet18_e5_va0.48108"),
auto_select=False,
batch_size=config.args.batch_size,
checkpoints_folder=config.args.checkpoints_folder,
classes=config.classes,
device=config.device,
eval_model=Path("checkpoints/resnet18_e5_va0.48108.pt"),
num_classes=config.num_classes,
num_layers=config.args.num_layers,
num_workers=config.args.num_workers,
path_mean=config.path_mean,
path_std=config.path_std,
pretrain=config.args.pretrain)
get_predictions(patches_eval_folder=Path(
"/home/ifsdata/vlg/jason/easy-self-supervised-training/data/voc_trainval_full/val"
),
output_folder=Path("outputs").joinpath(
"resnet18_e10_va0.55588"),
auto_select=False,
# DeepSlide
# Jason Wei, Behnaz Abdollahi, Saeed Hassanpour
# Run the resnet on generated patches.
from utils_model import get_predictions, get_predictions
# validation patches
get_predictions(patches_eval_folder=config.patches_eval_val,
auto_select=config.auto_select,
eval_model=config.eval_model,
checkpoints_folder=config.checkpoints_folder,
output_folder=config.preds_val)
# test patches
get_predictions(patches_eval_folder=config.patches_eval_test,
auto_select=config.auto_select,
eval_model=config.eval_model,
checkpoints_folder=config.checkpoints_folder,
output_folder=config.preds_test)
import config
from pathlib import Path
from utils_model import get_predictions
# Run the ResNet on the generated patches.
print("\n\n+++++ Running 4_test.py +++++")
print("\n----- Finding validation patch predictions -----")
# Validation patches.
get_predictions(
patches_eval_folder=Path(
"/home/brenta/scratch/data/imagenet_rotnet/train"),
#patches_eval_folder=Path("/home/ifsdata/vlg/jason/easy-self-supervised-training/data/voc_trainval_full/train"),
output_folder=Path("/home/brenta/scratch/jason/outputs/image_net/vanilla/"
).joinpath("resnet18_e0_mb40000_va0.80428.pt"),
auto_select=False,
batch_size=config.args.batch_size,
checkpoints_folder=config.args.checkpoints_folder,
classes=config.classes,
device=config.device,
eval_model=Path(
"/home/brenta/scratch/jason/checkpoints/image_net/vanilla/exp_10/resnet18_e0_mb40000_va0.80428.pt"
),
num_classes=config.num_classes,
num_layers=config.args.num_layers,
num_workers=config.args.num_workers,
path_mean=config.path_mean,
path_std=config.path_std,
pretrain=config.args.pretrain)
print("----- Finished finding validation patch predictions -----\n")
num_epochs=config.args.num_epochs,
train_folder=config.args.train_folder,
weight_decay=config.args.weight_decay)
print("+++++ Finished running 3_train.py +++++\n\n")
# Run the ResNet on the generated patches.
print("\n\n+++++ Running 4_test.py +++++")
print("\n----- Finding validation patch predictions -----")
# Validation patches.
get_predictions(patches_eval_folder=config.args.patches_eval_val,
output_folder=config.args.preds_val,
auto_select=config.args.auto_select,
batch_size=config.args.batch_size,
checkpoints_folder=config.args.checkpoints_folder,
classes=config.classes,
device=config.device,
eval_model=config.eval_model,
num_classes=config.num_classes,
num_layers=config.args.num_layers,
num_workers=config.args.num_workers,
path_mean=config.path_mean,
path_std=config.path_std,
pretrain=config.args.pretrain)
print("----- Finished finding validation patch predictions -----\n")
print("----- Finding test patch predictions -----")
# Test patches.
get_predictions(patches_eval_folder=config.args.patches_eval_test,
output_folder=config.args.preds_test,
auto_select=config.args.auto_select,
batch_size=config.args.batch_size,
checkpoints_folder=config.args.checkpoints_folder,
classes=config.classes,