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)
ccnnv9 leaky relu	5fd462122ef62498a51dd83db2c89507254d7a96	Thai Thien	2020-06-13 04:40:41
CompactCNNv9 leaky relu	03e5f342cc6eb3ba4fd0b98061398acedde6ab54	Thai Thien	2020-06-13 04:36:43
train script	ad55dbff34ffab8cffd068536c4d2f2b57a487ee	Thai Thien	2020-06-13 04:17:21
add L1Mean, add experiment tag on cometml	125f03f6f136bde9fa11b00a749b8a0c3de51534	Thai Thien	2020-06-13 04:15:30
msenone	f68f48cc1a2e28a65da368a9a66ebec9f8694467	Thai Thien	2020-06-13 03:54:44
train script	2b142c16cce5403d0d673077d2b1e442af072e42	Thai Thien	2020-06-12 16:42:25
train script	67923fc98cd4f82880e973f7dbf143c09ffcf8e8	Thai Thien	2020-06-12 16:01:59
ccnnv8 adam t1 sha, H3	0cab419a2c9747611347b6ddb75ca8a5cb20eb79	Thai Thien	2020-06-11 16:21:47
ccnv8 adam t1 shb	edd704f19c3d58d959c13b96100498c8809b309c	Thai Thien	2020-06-11 15:51:21
ccnnv8	4e8e6e864c3f2fea80a3b61addbc26eb8f3ae059	Thai Thien	2020-06-11 15:47:04
bigtail7 cuda4	fdde3eb95706e95cdc0e8d81f93f2e08036d3192	Thai Thien	2020-06-11 15:45:38
train script, bigtail6, bigtail7 shb	ea5758dcd7d8428e694cc84d50c81267eabcef23	Thai Thien	2020-06-11 15:36:54
WIP	3d7e1f10700c05bf7672cbf46ae6cafae8f2330f	Thai Thien	2020-06-11 13:07:18
return x	c750d65fe266979e9cdebf9a53b2af343462fbda	Thai Thien	2020-06-11 12:39:13
bigtail5	b750f0037993fdfdda6433b1982f5612d13aa1a6	Thai Thien	2020-06-11 12:32:28
adam t3 t4	933528ca8870affc741e989520172819382db0c0	Thai Thien	2020-06-10 07:59:26
adam t1 and adam t2	4202f1b609dc90197990e014addce68a04d0b8a8	Thai Thien	2020-06-09 15:23:49
t8	2dda6cd71a7aa50e760dfdcdfbdcdedfce36b836	Thai Thien	2020-06-09 15:12:20
t7	2ff8d0931ae51a93915721202aaa17c38c2918cc	Thai Thien	2020-06-09 15:08:32
t5 t6	7eeeb38e47fc4b59750438d66c911a7a93573b91	Thai Thien	2020-06-09 14:53:05

Commit 5fd462122ef62498a51dd83db2c89507254d7a96 - ccnnv9 leaky relu
Author: Thai Thien
Author date (UTC): 2020-06-13 04:40
Committer name: Thai Thien
Committer date (UTC): 2020-06-13 04:40
Parent(s): 03e5f342cc6eb3ba4fd0b98061398acedde6ab54
Signer:
Signing key:
Signing status: N
Tree: 3c755b91de5b7693e9bc4c526a4e71aae5ed5a7d

File	Lines added	Lines deleted
models/compact_cnn.py	46	0
train_script/adam_stuff/ccnnv9_adam_t1_sha.sh	6	4
train_script/adam_stuff/ccnnv9_adam_t2_sha.sh	6	6

File models/compact_cnn.py changed (mode: 100644) (index f8e9e2b..2976589)
...	...	class CompactCNNV7(nn.Module):
134	134	return x	return x
135	135
136	136
	137		class CompactCNNV9(nn.Module):
	138		"""
	139		A REAL-TIME DEEP NETWORK FOR CROWD COUNTING
	140		https://arxiv.org/pdf/2002.06515.pdf
	141		"""
	142		def __init__(self, load_weights=False):
	143		super(CompactCNNV8, self).__init__()
	144		self.model_note = "CCNNv7 but use leaky relu"
	145		self.red_cnn = nn.Conv2d(3, 10, 9, padding=4)
	146		self.green_cnn = nn.Conv2d(3, 14, 7, padding=3)
	147		self.blue_cnn = nn.Conv2d(3, 16, 5, padding=2)
	148		self.c0 = nn.Conv2d(40, 40, 3, padding=1)
	149
	150		self.max_pooling = nn.MaxPool2d(kernel_size=2, stride=2)
	151
	152		self.c1 = nn.Conv2d(40, 60, 3, padding=1)
	153		self.c2 = nn.Conv2d(60, 40, 3, padding=1)
	154		self.c3 = nn.Conv2d(40, 20, 3, padding=1)
	155		self.c4 = nn.Conv2d(20, 10, 3, padding=1)
	156		self.output = nn.Conv2d(10, 1, 1)
	157
	158		def forward(self,x):
	159		x_red = F.leaky_relu(self.red_cnn(x), inplace=True)
	160		x_green = F.leaky_relu(self.green_cnn(x), inplace=True)
	161		x_blue = F.leaky_relu(self.blue_cnn(x), inplace=True)
	162
	163		x = torch.cat((x_red, x_green, x_blue), 1)
	164
	165		x = self.max_pooling(x)
	166		x = F.leaky_relu(self.c0(x), inplace=True)
	167
	168		x = F.leaky_relu(self.c1(x), inplace=True)
	169
	170		x = F.leaky_relu(self.c2(x), inplace=True)
	171		x = self.max_pooling(x)
	172
	173		x = F.leaky_relu(self.c3(x), inplace=True)
	174		x = self.max_pooling(x)
	175
	176		x = F.leaky_relu(self.c4(x), inplace=True)
	177
	178		x = self.output(x)
	179		return x
	180
	181
	182
137	183	class CompactCNNV8(nn.Module):	class CompactCNNV8(nn.Module):
138	184	"""	"""
139	185	A REAL-TIME DEEP NETWORK FOR CROWD COUNTING	A REAL-TIME DEEP NETWORK FOR CROWD COUNTING

File train_script/adam_stuff/ccnnv9_adam_t1_sha.sh copied from file train_script/meow_one/big_tail/bigtail3_t9_sha.sh (similarity 73%) (mode: 100644) (index e3a7ddb..ce78088)
1		task="bigtail3_t9_sha"
	1		task="ccnnv9_adam_t1_sha.sh"
2	2
3	3	CUDA_VISIBLE_DEVICES=4 OMP_NUM_THREADS=2 PYTHONWARNINGS="ignore" HTTPS_PROXY="http://10.60.28.99:86" nohup python experiment_main.py \	CUDA_VISIBLE_DEVICES=4 OMP_NUM_THREADS=2 PYTHONWARNINGS="ignore" HTTPS_PROXY="http://10.60.28.99:86" nohup python experiment_main.py \
4	4	--task_id $task \	--task_id $task \
5		--note "bigtail3 L1 hope better than baseline" \
6		--model "BigTail3" \
	5		--note "leaky relu" \
	6		--model "CompactCNNV8" \
7	7	--input /data/rnd/thient/thient_data/ShanghaiTech/part_A \	--input /data/rnd/thient/thient_data/ShanghaiTech/part_A \
8	8	--lr 1e-4 \	--lr 1e-4 \
9	9	--decay 1e-4 \	--decay 1e-4 \
10		--loss_fn "L1" \
	10		--loss_fn "MSEMean" \
11	11	--skip_train_eval \	--skip_train_eval \
	12		--batch_size 1 \
	13		--optim "adam" \
12	14	--datasetname shanghaitech_crop_random \	--datasetname shanghaitech_crop_random \
13	15	--epochs 1200 > logs/$task.log &	--epochs 1200 > logs/$task.log &
14	16

File train_script/adam_stuff/ccnnv9_adam_t2_sha.sh copied from file train_script/adam_stuff/ccnn_adam_t4_sha.sh (similarity 74%) (mode: 100644) (index 34cb92f..53bb1f6)
1		task="ccnn_adam_t4_sha"
	1		task="ccnnv9_adam_t2_sha.sh"
2	2
3	3	CUDA_VISIBLE_DEVICES=5 OMP_NUM_THREADS=2 PYTHONWARNINGS="ignore" HTTPS_PROXY="http://10.60.28.99:86" nohup python experiment_main.py \	CUDA_VISIBLE_DEVICES=5 OMP_NUM_THREADS=2 PYTHONWARNINGS="ignore" HTTPS_PROXY="http://10.60.28.99:86" nohup python experiment_main.py \
4	4	--task_id $task \	--task_id $task \
5		--note "adam lr and decay, 8" \
6		--model "CompactCNNV7" \
	5		--note "leaky relu" \
	6		--model "CompactCNNV8" \
7	7	--input /data/rnd/thient/thient_data/ShanghaiTech/part_A \	--input /data/rnd/thient/thient_data/ShanghaiTech/part_A \
8		--lr 1e-6 \
9		--decay 1e-6 \
10		--loss_fn "MSEMean" \
	8		--lr 1e-4 \
	9		--decay 1e-4 \
	10		--loss_fn "L1Mean" \
11	11	--skip_train_eval \	--skip_train_eval \
12	12	--batch_size 1 \	--batch_size 1 \
13	13	--optim "adam" \	--optim "adam" \

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