RocketGit

hahattpro / crowd_counting_framework (public) (License: Unspecified) (since 2020-12-23) (hash sha1)

backupfrom https://github.com/ttpro1995/crowd_counting_framework

Clone URLs: https://rocketgit.com/user/hahattpro/crowd_counting_framework ssh://rocketgit@ssh.rocketgit.com/user/hahattpro/crowd_counting_framework git://git.rocketgit.com/user/hahattpro/crowd_counting_framework

List of commits:

Subject	Hash	Author	Date (UTC)
data_flow shanghaitech_pacnn_with_perspective seem working	91d350a06f358e03223966297d124daee94123d0	Thai Thien	2019-10-06 01:31:11
multiscale loss and final loss only mode	c65dd0e74ad28503821e5c8651a3b47b4a0c7c64	Thai Thien	2019-10-05 15:58:19
wip : perspective map	eac63f2671dc5b064753acc4f40bf0f9f216ad2a	Thai Thien	2019-10-04 16:26:56
shell script	f2106e700b6f6174d4dd276f25ec6f3d9ff239bb	thient	2019-10-04 07:42:51
WIP	42c7c8e1d772fbbda61a4bdf9e329f74e1efb600	tthien	2019-10-03 17:52:47
add readme	580cf43d1edddd67b1f6a2c57fdd5cee3dba925c	Thai Thien	2019-10-02 17:44:49
update script, debug	ddb68b95389be1c1d398118677dd227a8bb2b70b	Thai Thien	2019-10-02 15:52:31
add d (output density map) to loss function)	a0c71bf4bf2ab7393d60b06a84db8dfbbfb1a6c2	tthien	2019-09-30 16:32:39
fix the args, add save interval for model, so we don't save them all	9fdf9daa2ac4bd12b7b62521d81e520db0debd01	tthien	2019-09-30 16:30:00
meow	1ad19a22a310992e27a26471feeb37375124d075	tthien	2019-09-29 18:25:43
fix pacnn perspective map	453ece3ccb818889ba895bfc4285f7905d33cba5	Thai Thien	2019-09-25 17:20:33
apex not work so well	da8c0dd57297f972201f31d57e66897177922f48	Thai Thien	2019-09-24 17:25:59
fix data loader pacnn so it will scale up with correct number of people	11d55b50d764511f2491291f0208fee0905dec49	Thai Thien	2019-09-24 15:40:56
add comet ml	a9d4b89ce594f5e241168ccafdcdf0f150ea0ebb	Thai Thien	2019-09-23 17:07:58
fix pacnn avg schema	c2140a96886195782e5689c24aeeb4fe7a2db7ad	Thai Thien	2019-09-22 17:35:01
debug number not divisible by 8	a568fd7f294a8bd31b3db78437b4b6b51b5b41b9	Thai Thien	2019-09-22 04:36:06
pacnn	967074890d14ab0eefc277801860270a468e8f9f	Thai Thien	2019-09-22 03:54:48
wip: pacnn	2192d7c7b449fecf3868877d9cfbc09bb6f7ae98	Thai Thien	2019-09-22 03:44:56
wip: pacnn	37620e5a9bc0f9516ea964ec58d9bdaa1c40ff36	Thai Thien	2019-09-22 03:14:42
fix training flow	2b87b1b26c7296b64493fdc49fedb421b249dfa3	Thai Thien	2019-09-17 18:00:35

Commit 91d350a06f358e03223966297d124daee94123d0 - data_flow shanghaitech_pacnn_with_perspective seem working
Author: Thai Thien
Author date (UTC): 2019-10-06 01:31
Committer name: Thai Thien
Committer date (UTC): 2019-10-06 01:31
Parent(s): c65dd0e74ad28503821e5c8651a3b47b4a0c7c64
Signing key:
Tree: 2d619b8ad96bbeae323dcc638a9e794f932051de

File	Lines added	Lines deleted
data_flow.py	9	6
hard_code_variable.py	2	1
train_script/train_pacnn_shanghaitechA.sh	10	1
visualize_data_loader.py	42	1

File data_flow.py changed (mode: 100644) (index 19347e3..a7b2270)
...	...	def load_data_shanghaitech_pacnn_with_perspective(img_path, train=True):
114	114	:return:	:return:
115	115	"""	"""
116	116	gt_path = img_path.replace('.jpg', '.h5').replace('images', 'ground-truth-h5')	gt_path = img_path.replace('.jpg', '.h5').replace('images', 'ground-truth-h5')
117		p_path = img_path.replace(".jpg", ".mat").replace("images", "p_map")
	117		p_path = img_path.replace(".jpg", ".mat").replace("images", "pmap")
118	118	img = Image.open(img_path).convert('RGB')	img = Image.open(img_path).convert('RGB')
119	119	gt_file = h5py.File(gt_path, 'r')	gt_file = h5py.File(gt_path, 'r')
120	120	target = np.asarray(gt_file['density'])	target = np.asarray(gt_file['density'])
121		perspective = np.array(h5py.File(p_path, "r"))
122
	121		perspective = np.array(h5py.File(p_path, "r")['pmap'])
	122		perspective = np.rot90(perspective, k=3)
123	123	if train:	if train:
124	124	crop_size = (int(img.size[0] / 2), int(img.size[1] / 2))	crop_size = (int(img.size[0] / 2), int(img.size[1] / 2))
125	125	if random.randint(0, 9) <= -1:	if random.randint(0, 9) <= -1:

...	...	def load_data_shanghaitech_pacnn_with_perspective(img_path, train=True):
132	132
133	133	img = img.crop((dx, dy, crop_size[0] + dx, crop_size[1] + dy))	img = img.crop((dx, dy, crop_size[0] + dx, crop_size[1] + dy))
134	134	target = target[dy:crop_size[1] + dy, dx:crop_size[0] + dx]	target = target[dy:crop_size[1] + dy, dx:crop_size[0] + dx]
135		perspective = target[dy:crop_size[1] + dy, dx:crop_size[0] + dx]
	135		perspective = perspective[dy:crop_size[1] + dy, dx:crop_size[0] + dx]
136	136	if random.random() > 0.8:	if random.random() > 0.8:
137	137	target = np.fliplr(target)	target = np.fliplr(target)
138	138	img = img.transpose(Image.FLIP_LEFT_RIGHT)	img = img.transpose(Image.FLIP_LEFT_RIGHT)

...	...	def load_data_shanghaitech_pacnn_with_perspective(img_path, train=True):
148	148	perspective_s = cv2.resize(perspective, (int(perspective.shape[1] / 16), int(perspective.shape[0] / 16)),	perspective_s = cv2.resize(perspective, (int(perspective.shape[1] / 16), int(perspective.shape[0] / 16)),
149	149	interpolation=cv2.INTER_CUBIC) * 256	interpolation=cv2.INTER_CUBIC) * 256
150	150
151		perspective_m = cv2.resize(perspective, (int(perspective.shape[1] / 8), int(perspective.shape[0] / 8)),
	151		perspective_p = cv2.resize(perspective, (int(perspective.shape[1] / 8), int(perspective.shape[0] / 8)),
152	152	interpolation=cv2.INTER_CUBIC) * 64	interpolation=cv2.INTER_CUBIC) * 64
153	153
154		return img, (target1, target2, target3, perspective_s, perspective_m)
	154		return img, (target1, target2, target3, perspective_s, perspective_p)
	155
155	156
156	157	def load_data_ucf_cc50_pacnn(img_path, train=True):	def load_data_ucf_cc50_pacnn(img_path, train=True):
157	158	"""	"""

...	...	class ListDataset(Dataset):
258	259	self.load_data_fn = load_data_ucf_cc50_pacnn	self.load_data_fn = load_data_ucf_cc50_pacnn
259	260	elif dataset_name is "shanghaitech_pacnn":	elif dataset_name is "shanghaitech_pacnn":
260	261	self.load_data_fn = load_data_shanghaitech_pacnn	self.load_data_fn = load_data_shanghaitech_pacnn
	262		elif dataset_name is "shanghaitech_pacnn_with_perspective":
	263		self.load_data_fn = load_data_shanghaitech_pacnn_with_perspective
261	264
262	265	def __len__(self):	def __len__(self):
263	266	return self.nSamples	return self.nSamples

File hard_code_variable.py changed (mode: 100644) (index 836144a..3f5b17a)
1	1	class HardCodeVariable():	class HardCodeVariable():
2	2	def __init__(self):	def __init__(self):
3		self.UCF_CC_50_PATH = "/data/cv_data/UCFCrowdCountingDataset_CVPR13_with_people_density_map/UCF_CC_50"
	3		self.UCF_CC_50_PATH = "/data/cv_data/UCFCrowdCountingDataset_CVPR13_with_people_density_map/UCF_CC_50"
	4		self.SHANGHAITECH_PATH = "data/ShanghaiTech/part_A/train_data"

File train_script/train_pacnn_shanghaitechA.sh changed (mode: 100644) (index 9f2955d..f3944dd)
45	45	#--load_model saved_model/train_state1_attemp6_120_checkpoint.pth.tar \	#--load_model saved_model/train_state1_attemp6_120_checkpoint.pth.tar \
46	46	#--epochs 300 \	#--epochs 300 \
47	47	#--lr 1e-9 \	#--lr 1e-9 \
48		#--task_id train_state1_attemp7
	48		#--task_id train_state1_attemp7
	49
	50		#### no loss for d1, d2, d3 but only count d_final
	51		python main_pacnn.py \
	52		--input data/ShanghaiTech/part_A \
	53		--load_model saved_model/train_state1_attemp7_180_checkpoint.pth.tar \
	54		--epochs 300 \
	55		--lr 1e-9 \
	56		--PACNN_MUTILPLE_SCALE_LOSS False \
	57		--task_id train_state1_attemp8_finalloss

File visualize_data_loader.py changed (mode: 100644) (index b515bbd..f4aa432)
...	...	def visualize_ucf_cc_50_pacnn():
48	48	print("count2 ", label[1].numpy()[0].sum())	print("count2 ", label[1].numpy()[0].sum())
49	49	print("count3 ", label[2].numpy()[0].sum())	print("count3 ", label[2].numpy()[0].sum())
50	50
	51		def visualize_shanghaitech_pacnn_with_perspective():
	52		HARD_CODE = HardCodeVariable()
	53		saved_folder = "visualize/test_dataloader"
	54		os.makedirs(saved_folder, exist_ok=True)
	55		DATA_PATH = HARD_CODE.SHANGHAITECH_PATH
	56		train_list, val_list = get_train_val_list(DATA_PATH, test_size=0.2)
	57		test_list = None
	58
	59		# create data loader
	60		train_loader, val_loader, test_loader = get_dataloader(train_list, val_list, test_list, dataset_name="ucf_cc_50")
	61		train_loader_pacnn = torch.utils.data.DataLoader(
	62		ListDataset(train_list,
	63		shuffle=True,
	64		transform=transforms.Compose([
	65		transforms.ToTensor()
	66		]),
	67		train=True,
	68		batch_size=1,
	69		num_workers=4, dataset_name="shanghaitech_pacnn_with_perspective", debug=True),
	70		batch_size=1, num_workers=4)
	71
	72		img, label = next(iter(train_loader_pacnn))
	73
	74		print(img.shape)
	75		save_img(img, os.path.join(saved_folder, "pacnn_loader_img.png"))
	76		save_density_map(label[0].numpy()[0], os.path.join(saved_folder,"pacnn_loader_with_p_density1.png"))
	77		save_density_map(label[1].numpy()[0], os.path.join(saved_folder,"pacnn_loader_with_p_density2.png"))
	78		save_density_map(label[2].numpy()[0], os.path.join(saved_folder,"pacnn_loader_with_p_density3.png"))
	79		save_density_map(label[3].numpy()[0], os.path.join(saved_folder, "pacnn_loader_p_s_4.png"))
	80		save_density_map(label[4].numpy()[0], os.path.join(saved_folder, "pacnn_loader_p_5.png"))
	81		print("count1 ", label[0].numpy()[0].sum())
	82		print("count2 ", label[1].numpy()[0].sum())
	83		print("count3 ", label[2].numpy()[0].sum())
	84		print("count4 ", label[3].numpy()[0].sum())
	85		print("count5 ", label[4].numpy()[0].sum())
	86
	87		print("s1 ", label[0].shape)
	88		print("s2 ", label[1].shape)
	89		print("s3 ", label[2].shape)
	90		print("s4 ", label[3].shape)
	91		print("s5 ", label[4].shape)
51	92
52	93	if __name__ == "__main__":	if __name__ == "__main__":
53		visualize_ucf_cc_50_pacnn()
	94		visualize_shanghaitech_pacnn_with_perspective()

Hints:
Before first commit, do not forget to setup your git environment:

git config --global user.name "your_name_here"
git config --global user.email "your@email_here"

Clone this repository using HTTP(S):

git clone https://rocketgit.com/user/hahattpro/crowd_counting_framework

Clone this repository using ssh (do not forget to upload a key first):

git clone ssh://rocketgit@ssh.rocketgit.com/user/hahattpro/crowd_counting_framework

Clone this repository using git:

git clone git://git.rocketgit.com/user/hahattpro/crowd_counting_framework

You are allowed to anonymously push to this repository.
This means that your pushed commits will automatically be transformed into a merge request:

... clone the repository ...
... make some changes and some commits ...
git push origin main