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)
mse-ssim loss function	052de7c879bff7690a7cfc1905c8376bf8605c45	Thai Thien	2020-04-30 11:41:24
delete misinformation note	ca07533d585def04eec086685f2e72eacc330ddc	Thai Thien	2020-04-29 16:59:47
typo	3505077301af5349665f12121862db2512ad450a	Thai Thien	2020-04-29 16:59:03
change, adapt, survive	6755c8375302af2146cb63adc967631d53a7b1c8	Thai Thien	2020-04-29 16:48:10
gpu	861ec0d41cea2eec359c6ddfe207e1ed6b583369	Thai Thien	2020-04-28 18:08:11
ccnn_v7_t4	a84f63e64fe8b31fe22c94d383de5ed4e1a27fe4	Thai Thien	2020-04-28 18:07:33
2500	d37516f53d21f3dfc05143ba5ffda80fc5a07825	Thai Thien	2020-04-28 17:56:57
change max epoch	246ae30e6dd2b4a15b7dc70a3dc05592ac1c48f2	Thai Thien	2020-04-28 17:55:23
h1 t3 h1 t4	fbcd13dd240e06a982a1ce48f27cd1d542a26b63	Thai Thien	2020-04-28 17:53:06
h1	96204cb5262500020371637741131d24e3fa91d0	Thai Thien	2020-04-27 17:35:51
typo	adfb213c2564bc90b8b69811469534b004808644	Thai Thien	2020-04-27 17:17:58
batch size 8 for shb	c90fa9a5d725a1ef0d29ed23f947ee05b9aa7894	Thai Thien	2020-04-27 17:10:34
change proxy	30cd53782eb17b416c471502f1e6c6e9975a644b	Thai Thien	2020-04-27 17:06:28
experiment with CCNN when we max pooling after merge	c8e8daec89ee71c2f30e30cf3030298ee7073e56	Thai Thien	2020-04-27 16:52:50
refactor train_compact_cnn	9d1ecece2382b79a98e6cf2d4579ae68172dbb6a	Thai Thien	2020-04-27 16:35:47
we add head experiment	82a25e6b89945609486cbafc433eaea20dcdee39	Thai Thien	2020-04-27 16:23:50
CompactCNNV3	7de3766d085ebdbdf82b024eb517568dd82d8d6d	Thai Thien	2020-04-27 16:23:20
no_norm	da3c84dca19b0d281082679d88af3b9d27165bfe	Thai Thien	2020-04-25 17:32:45
M4_t3_sha_c_shb	d0d61ff74ed23f595d05d6a813c0a93239f61438	Thai Thien	2020-04-25 17:17:56
training script	624ecec7b12641f734e12ee2ebb6158c7c89683a	Thai Thien	2020-04-25 17:08:25

Commit 052de7c879bff7690a7cfc1905c8376bf8605c45 - mse-ssim loss function
Author: Thai Thien
Author date (UTC): 2020-04-30 11:41
Committer name: Thai Thien
Committer date (UTC): 2020-04-30 11:41
Parent(s): ca07533d585def04eec086685f2e72eacc330ddc
Signer:
Signing key:
Signing status: N
Tree: 39310bbb32b85d8c65523e68577ff99f812b4a50

File	Lines added	Lines deleted
args_util.py	2	0
local_train_script/ccnn_t2_shb.sh	12	0
mse_ssim_loss.py	13	0
pytorch_ssim/README.md	2	0
train_compact_cnn.py	9	4

File args_util.py changed (mode: 100644) (index d83740e..c646aeb)
...	...	def meow_parse():
122	122	parser.add_argument('--test', action="store_true", default=False)	parser.add_argument('--test', action="store_true", default=False)
123	123	parser.add_argument('--no_norm', action="store_true", default=False,	parser.add_argument('--no_norm', action="store_true", default=False,
124	124	help="if true, does not use transforms.Normalize in dataloader")	help="if true, does not use transforms.Normalize in dataloader")
	125		parser.add_argument('--use_ssim', action="store_true", default=False,
	126		help="if true, use mse and negative ssim as loss function")
125	127	arg = parser.parse_args()	arg = parser.parse_args()
126	128	return arg	return arg
127	129

File local_train_script/ccnn_t2_shb.sh added (mode: 100644) (index 0000000..16ab5e7)
	1	task="local_ccnn_t2_shb"
	2
	3	python train_compact_cnn.py \
	4	--task_id $task \
	5	--note "" \
	6	--model "CompactCNNV7" \
	7	--input /data/ShanghaiTech/part_B \
	8	--lr 1e-4 \
	9	--decay 1e-4 \
	10	--batch_size 8 \
	11	--datasetname shanghaitech_rnd \
	12	--epochs 301

File mse_ssim_loss.py added (mode: 100644) (index 0000000..babe708)
	1	import torch
	2	from torch import nn
	3	from pytorch_ssim import SSIM
	4
	5
	6	class MseSsimLoss(torch.nn.Module):
	7	def __init__(self):
	8	super(MseSsimLoss, self).__init__()
	9	self.mse = nn.MSELoss(reduction='sum')
	10	self.ssim = SSIM(window_size=5)
	11
	12	def forward(self, input, target):
	13	return self.mse(input, target) - self.ssim(input, target)

File pytorch_ssim/README.md added (mode: 100644) (index 0000000..8cb0a17)
	1	Here is original github
	2	https://github.com/Po-Hsun-Su/pytorch-ssim

File train_compact_cnn.py changed (mode: 100644) (index 7b57017..4b570dc)
...	...	from ignite.metrics import Loss
7	7	from ignite.handlers import Checkpoint, DiskSaver, Timer	from ignite.handlers import Checkpoint, DiskSaver, Timer
8	8	from crowd_counting_error_metrics import CrowdCountingMeanAbsoluteError, CrowdCountingMeanSquaredError	from crowd_counting_error_metrics import CrowdCountingMeanAbsoluteError, CrowdCountingMeanSquaredError
9	9	from visualize_util import get_readable_time	from visualize_util import get_readable_time
10
	10		from mse_ssim_loss import MseSsimLoss
11	11	import torch	import torch
12	12	from torch import nn	from torch import nn
	13		from pytorch_ssim import SSIM
13	14
14	15	from models import CompactCNNV2, CompactCNNV7	from models import CompactCNNV2, CompactCNNV7
15	16

...	...	import os
17	18	from model_util import get_lr	from model_util import get_lr
18	19
19	20	COMET_ML_API = "S3mM1eMq6NumMxk2QJAXASkUM"	COMET_ML_API = "S3mM1eMq6NumMxk2QJAXASkUM"
20		PROJECT_NAME = "crowd-counting-framework"
21		# PROJECT_NAME = "crowd-counting-debug"
	21		# PROJECT_NAME = "crowd-counting-framework"
	22		PROJECT_NAME = "crowd-counting-debug"
22	23
23	24
24	25	def very_simple_param_count(model):	def very_simple_param_count(model):

...	...	if __name__ == "__main__":
80	81	model = model.to(device)	model = model.to(device)
81	82
82	83	# loss function	# loss function
83		loss_fn = nn.MSELoss(reduction='sum').to(device)
	84		if args.use_ssim:
	85		loss_fn = MseSsimLoss()
	86		print("use ssim")
	87		else:
	88		loss_fn = nn.MSELoss(reduction='sum').to(device)
84	89
85	90	optimizer = torch.optim.Adam(model.parameters(), args.lr,	optimizer = torch.optim.Adam(model.parameters(), args.lr,
86	91	weight_decay=args.decay)	weight_decay=args.decay)

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