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)
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
evaluator	ffc5bf8290ae0c469a9a18a2d061cfd1bfeee822	Thai Thien	2019-09-14 04:56:35
some more test for data loader	25173578cde7d4e9fe6c6140d1ee01caa4fcfc32	Thai Thien	2019-09-14 02:51:58
some visualize to debug data loader	e4f52007616acf307bddbde79c0fb4f8c649c785	Thai Thien	2019-09-13 17:35:45
wip	d7d44cad6774355bdfa45414258763f6c6a0c299	Thai Thien	2019-08-31 16:58:16
commit all	6dad7a58f7dbf9fc288ce9dd3e92be538851c2a7	Thai Thien	2019-08-29 19:10:44
input d1,d2,d3 match	fc2a809241f8b6356d964c63d40cbebd55ca5f6c	Thai Thien	2019-08-28 17:57:05
WIP	39eab26d061e61dfffbf164dbd5fd878299b7250	thient	2019-08-28 11:09:12
output of de is ok	dd770386674df3e0fbebafdfc48a9352bc28967d	thient	2019-08-28 10:54:09

Commit 1ad19a22a310992e27a26471feeb37375124d075 - meow
Author: tthien
Author date (UTC): 2019-09-29 18:25
Committer name: tthien
Committer date (UTC): 2019-09-29 18:25
Parent(s): 453ece3ccb818889ba895bfc4285f7905d33cba5
Signing key:
Tree: c011d03d1eb4b527115e109816a40e6d502cc486

File	Lines added	Lines deleted
main_pacnn.py	25	16
model_util.py	8	2
train_script/train_pacnn_shanghaitechA.sh	1	1

File main_pacnn.py changed (mode: 100644) (index 43b6c3d..b42c5a0)
1	1	from comet_ml import Experiment	from comet_ml import Experiment
2	2	from args_util import real_args_parse	from args_util import real_args_parse
3	3	from data_flow import get_train_val_list, get_dataloader, create_training_image_list	from data_flow import get_train_val_list, get_dataloader, create_training_image_list
4		from ignite.engine import Events, create_supervised_trainer, create_supervised_evaluator
5		from ignite.metrics import Loss, MeanAbsoluteError, MeanSquaredError
6	4	from crowd_counting_error_metrics import CrowdCountingMeanAbsoluteError, CrowdCountingMeanSquaredError	from crowd_counting_error_metrics import CrowdCountingMeanAbsoluteError, CrowdCountingMeanSquaredError
7	5	import torch	import torch
8	6	from torch import nn	from torch import nn

...	...	from evaluator import MAECalculator
18	16
19	17	from model_util import save_checkpoint	from model_util import save_checkpoint
20	18
21		import apex
22		from apex import amp
	19		# import apex
	20		# from apex import amp
23	21
24	22	if __name__ == "__main__":	if __name__ == "__main__":
25	23	# import comet_ml in the top of your file	# import comet_ml in the top of your file
26	24
27	25
28		MODEL_SAVE_NAME = "dev5"
	26		MODEL_SAVE_NAME = "dev7"
29	27	# Add the following code anywhere in your machine learning file	# Add the following code anywhere in your machine learning file
30	28	experiment = Experiment(api_key="S3mM1eMq6NumMxk2QJAXASkUM",	experiment = Experiment(api_key="S3mM1eMq6NumMxk2QJAXASkUM",
31	29	project_name="pacnn-dev2", workspace="ttpro1995")	project_name="pacnn-dev2", workspace="ttpro1995")

...	...	if __name__ == "__main__":
40	38	print(args)	print(args)
41	39	DATA_PATH = args.input	DATA_PATH = args.input
42	40	DATASET_NAME = "shanghaitech"	DATASET_NAME = "shanghaitech"
	41		TOTAL_EPOCH = args.epochs
43	42	PACNN_PERSPECTIVE_AWARE_MODEL = True	PACNN_PERSPECTIVE_AWARE_MODEL = True
44	43
45	44

...	...	if __name__ == "__main__":
89	88	momentum=args.momentum,	momentum=args.momentum,
90	89	weight_decay=args.decay)	weight_decay=args.decay)
91	90	# Allow Amp to perform casts as required by the opt_level	# Allow Amp to perform casts as required by the opt_level
92		net, optimizer = amp.initialize(net, optimizer, opt_level="O1", enabled=False)
93
94		for e in range(10):
95		print("start epoch ", e)
	91		# net, optimizer = amp.initialize(net, optimizer, opt_level="O1", enabled=False)
	92
	93		current_save_model_name = ""
	94		current_epoch = 0
	95		while (current_epoch < TOTAL_EPOCH):
	96		experiment.log_current_epoch(current_epoch)
	97		current_epoch += 1
	98		print("start epoch ", current_epoch)
96	99	loss_sum = 0	loss_sum = 0
97	100	sample = 0	sample = 0
98	101	start_time = time()	start_time = time()

...	...	if __name__ == "__main__":
120	123	pass	pass
121	124	loss_d = criterion_mse(d, d1_label) + criterion_ssim(d, d1_label)	loss_d = criterion_mse(d, d1_label) + criterion_ssim(d, d1_label)
122	125	loss += loss_d	loss += loss_d
123		# loss.backward()
124		with amp.scale_loss(loss, optimizer) as scaled_loss:
125		scaled_loss.backward()
	126		loss.backward()
	127		# with amp.scale_loss(loss, optimizer) as scaled_loss:
	128		# scaled_loss.backward()
126	129	optimizer.step()	optimizer.step()
127	130	optimizer.zero_grad()	optimizer.zero_grad()
128	131	loss_sum += loss.item()	loss_sum += loss.item()

...	...	if __name__ == "__main__":
135	138	experiment.log_metric("avg_loss_ministep", avg_loss_ministep)	experiment.log_metric("avg_loss_ministep", avg_loss_ministep)
136	139	# if counting == 100:	# if counting == 100:
137	140	# break	# break
	141
138	142	end_time = time()	end_time = time()
139	143	avg_loss = loss_sum/sample	avg_loss = loss_sum/sample
140	144	epoch_time = end_time - start_time	epoch_time = end_time - start_time
141		print("==END epoch ", e, " =============================================")
	145		print("==END epoch ", current_epoch, " =============================================")
142	146	print(epoch_time, avg_loss, sample)	print(epoch_time, avg_loss, sample)
143	147	experiment.log_metric("avg_loss_epoch", avg_loss)	experiment.log_metric("avg_loss_epoch", avg_loss)
144	148	print("=================================================================")	print("=================================================================")
145	149
146		save_checkpoint({
	150		current_save_model_name = save_checkpoint({
147	151	'model': net.state_dict(),	'model': net.state_dict(),
148	152	'optimizer': optimizer.state_dict(),	'optimizer': optimizer.state_dict(),
	153		'e': current_epoch,
	154		'PACNN_PERSPECTIVE_AWARE_MODEL': PACNN_PERSPECTIVE_AWARE_MODEL
149	155	# 'amp': amp.state_dict()	# 'amp': amp.state_dict()
150		}, False, MODEL_SAVE_NAME)
	156		}, False, MODEL_SAVE_NAME+"_"+str(current_epoch)+"_")
	157
	158		experiment.log_asset(current_save_model_name)
151	159
	160		# end 1 epoch
152	161
153	162	# after epoch evaluate	# after epoch evaluate
154	163	mae_calculator_d1 = MAECalculator()	mae_calculator_d1 = MAECalculator()

...	...	if __name__ == "__main__":
189	198	net = PACNNWithPerspectiveMap(PACNN_PERSPECTIVE_AWARE_MODEL).to(device)	net = PACNNWithPerspectiveMap(PACNN_PERSPECTIVE_AWARE_MODEL).to(device)
190	199	print(net)	print(net)
191	200
192		best_checkpoint = torch.load(MODEL_SAVE_NAME + "checkpoint.pth.tar")
	201		best_checkpoint = torch.load(current_save_model_name)
193	202	net.load_state_dict(best_checkpoint['model'])	net.load_state_dict(best_checkpoint['model'])
194	203
195	204	# device = "cpu"	# device = "cpu"

File model_util.py changed (mode: 100644) (index 8bb883a..eeb3468)
...	...	import h5py
2	2	import torch	import torch
3	3	import shutil	import shutil
4	4	import numpy as np	import numpy as np
	5		import os
	6
5	7
6	8	def save_net(fname, net):	def save_net(fname, net):
7	9	with h5py.File(fname, 'w') as h5f:	with h5py.File(fname, 'w') as h5f:

...	...	def load_net(fname, net):
17	19
18	20
19	21	def save_checkpoint(state, is_best, task_id, filename='checkpoint.pth.tar'):	def save_checkpoint(state, is_best, task_id, filename='checkpoint.pth.tar'):
20		torch.save(state, task_id + filename)
	22		if not os.path.exists("saved_model"):
	23		os.makedirs("saved_model")
	24		full_file_name = os.path.join("saved_model", task_id + filename)
	25		torch.save(state, full_file_name)
21	26	if is_best:	if is_best:
22		shutil.copyfile(task_id + filename, task_id + 'model_best.pth.tar')
	27		shutil.copyfile(task_id + filename, task_id + 'model_best.pth.tar')
	28		return full_file_name

File train_script/train_pacnn_shanghaitechA.sh changed (mode: 100644) (index 34be054..ced1a20)
1	1	#python /home/tt/project/crowd_counting_framework/main_pacnn.py --input /home/tt/project/crowd_counting_framework/data/ShanghaiTech/part_A	#python /home/tt/project/crowd_counting_framework/main_pacnn.py --input /home/tt/project/crowd_counting_framework/data/ShanghaiTech/part_A
2	2
3		python /home/tt/project/crowd_counting_framework/main_pacnn.py --input /home/tt/project/crowd_counting_framework/data/ShanghaiTech/part_A
	3		python main_pacnn.py --input data/ShanghaiTech/part_A --epochs 3

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