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)
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
dataset script	bc5c052f5f956510ab95ef9a45434fd486c57fae	Thai Thien	2019-09-16 17:21:13

Commit c65dd0e74ad28503821e5c8651a3b47b4a0c7c64 - multiscale loss and final loss only mode
Author: Thai Thien
Author date (UTC): 2019-10-05 15:58
Committer name: Thai Thien
Committer date (UTC): 2019-10-05 15:58
Parent(s): eac63f2671dc5b064753acc4f40bf0f9f216ad2a
Signing key:
Tree: 51822798b2c4953ad1f02b3fb45bd2765a04a1b4

File	Lines added	Lines deleted
args_util.py	3	0
main_pacnn.py	13	6
train_script/train_pacnn_shanghaitechA.sh	24	6

File args_util.py changed (mode: 100644) (index c9c322c..fa52179)
...	...	def real_args_parse():
75	75
76	76	# pacnn setting only	# pacnn setting only
77	77	parser.add_argument('--PACNN_PERSPECTIVE_AWARE_MODEL', action="store_true", default=False)	parser.add_argument('--PACNN_PERSPECTIVE_AWARE_MODEL', action="store_true", default=False)
	78		parser.add_argument('--PACNN_MUTILPLE_SCALE_LOSS', action="store", default=True,
	79		help="True: compare each of density map/perspective map scale with gt for loss."
	80		" False: only compare final density map and final density perspective map")
78	81
79	82	# args.original_lr = 1e-7	# args.original_lr = 1e-7
80	83	# args.lr = 1e-7	# args.lr = 1e-7

File main_pacnn.py changed (mode: 100644) (index fc19585..fc87076)
...	...	if __name__ == "__main__":
37	37	DATASET_NAME = "shanghaitech"	DATASET_NAME = "shanghaitech"
38	38	TOTAL_EPOCH = args.epochs	TOTAL_EPOCH = args.epochs
39	39	PACNN_PERSPECTIVE_AWARE_MODEL = args.PACNN_PERSPECTIVE_AWARE_MODEL	PACNN_PERSPECTIVE_AWARE_MODEL = args.PACNN_PERSPECTIVE_AWARE_MODEL
	40		PACNN_MUTILPLE_SCALE_LOSS = args.PACNN_MUTILPLE_SCALE_LOSS
40	41
41	42	experiment.set_name(args.task_id)	experiment.set_name(args.task_id)
42	43	experiment.log_parameter("DATA_PATH", DATA_PATH)	experiment.log_parameter("DATA_PATH", DATA_PATH)
43	44	experiment.log_parameter("PACNN_PERSPECTIVE_AWARE_MODEL", PACNN_PERSPECTIVE_AWARE_MODEL)	experiment.log_parameter("PACNN_PERSPECTIVE_AWARE_MODEL", PACNN_PERSPECTIVE_AWARE_MODEL)
	45		experiment.log_parameter("PACNN_MUTILPLE_SCALE_LOSS", PACNN_MUTILPLE_SCALE_LOSS)
44	46	experiment.log_parameter("train", "train without p")	experiment.log_parameter("train", "train without p")
45	47	experiment.log_parameter("momentum", args.momentum)	experiment.log_parameter("momentum", args.momentum)
46	48	experiment.log_parameter("lr", args.lr)	experiment.log_parameter("lr", args.lr)

...	...	if __name__ == "__main__":
128	130	# forward pass	# forward pass
129	131
130	132	d1, d2, d3, p_s, p, d = net(train_img.to(device))	d1, d2, d3, p_s, p, d = net(train_img.to(device))
131		loss_1 = criterion_mse(d1, d1_label) + criterion_ssim(d1, d1_label)
132		loss_2 = criterion_mse(d2, d2_label) + criterion_ssim(d2, d2_label)
133		loss_3 = criterion_mse(d3, d3_label) + criterion_ssim(d3, d3_label)
134	133	loss_d = criterion_mse(d, d1_label) + criterion_ssim(d, d1_label)	loss_d = criterion_mse(d, d1_label) + criterion_ssim(d, d1_label)
135		loss = loss_d + loss_1 + loss_2 + loss_3
	134		loss = loss_d
	135
	136		if PACNN_MUTILPLE_SCALE_LOSS:
	137		loss_1 = criterion_mse(d1, d1_label) + criterion_ssim(d1, d1_label)
	138		loss_2 = criterion_mse(d2, d2_label) + criterion_ssim(d2, d2_label)
	139		loss_3 = criterion_mse(d3, d3_label) + criterion_ssim(d3, d3_label)
	140		loss += loss_1 + loss_2 + loss_3
	141
136	142
137	143	if PACNN_PERSPECTIVE_AWARE_MODEL:	if PACNN_PERSPECTIVE_AWARE_MODEL:
138	144	# TODO: loss for perspective map here	# TODO: loss for perspective map here
139	145	pass	pass
140		loss_d = criterion_mse(d, d1_label) + criterion_ssim(d, d1_label)
141		loss += loss_d
	146		# what is this, loss_d count 2 ?
	147		## loss_d = criterion_mse(d, d1_label) + criterion_ssim(d, d1_label)
	148		## loss += loss_d
142	149
143	150	# with amp.scale_loss(loss, optimizer) as scaled_loss:	# with amp.scale_loss(loss, optimizer) as scaled_loss:
144	151	# scaled_loss.backward()	# scaled_loss.backward()

File train_script/train_pacnn_shanghaitechA.sh changed (mode: 100644) (index 72ae080..9f2955d)
22	22	#--task_id train_state1_attemp4	#--task_id train_state1_attemp4
23	23
24	24
25		python main_pacnn.py \
26		--input data/ShanghaiTech/part_A \
27		--load_model saved_model/train_state1_attemp4_35_checkpoint.pth.tar \
28		--epochs 151 \
29		--lr 1e-8 \
30		--task_id train_state1_attemp5
	25		#python main_pacnn.py \
	26		#--input data/ShanghaiTech/part_A \
	27		#--load_model saved_model/train_state1_attemp4_35_checkpoint.pth.tar \
	28		#--epochs 151 \
	29		#--lr 1e-8 \
	30		#--task_id train_state1_attemp5
	31
	32		################3
	33
	34		## TODO: train this
	35		#python main_pacnn.py \
	36		#--input data/ShanghaiTech/part_A \
	37		#--load_model saved_model/train_state1_attemp5_40_checkpoint.pth.tar \
	38		#--epochs 300 \
	39		#--lr 1e-8 \
	40		#--task_id train_state1_attemp6
	41
	42
	43		#python main_pacnn.py \
	44		#--input data/ShanghaiTech/part_A \
	45		#--load_model saved_model/train_state1_attemp6_120_checkpoint.pth.tar \
	46		#--epochs 300 \
	47		#--lr 1e-9 \
	48		#--task_id train_state1_attemp7

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