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)
perfomance test	50d9652b43d660d108d697ad730e547aabf32e46	Thai Thien	2020-08-14 15:46:25
change name	dd91bf5002f1505d1dd214c13ea416dbb610e2e3	Thai Thien	2020-08-09 14:20:40
msemean	ede02b8197895eb82e24388a9bf59721736356a3	Thai Thien	2020-08-09 14:18:34
l2_adamw2_bigtail11i_t2_c3_sha.sh	ca44a2ea3f6cabb49738144fc3a918d3891280dc	Thai Thien	2020-08-09 14:09:38
l2_adamw2_bigtail11i_t2_c3_sha	3d69914ecd40cd8a9a4f9863fcaf0ed20fb293b9	Thai Thien	2020-08-09 14:03:56
l2_adamw2_bigtail11i_	4bc0b7938d27842c76fb9dd5fc68cc4eb66d9b73	Thai Thien	2020-08-09 13:57:45
c2	f74909d7d10e35b318fd64814304805055b7c927	Thai Thien	2020-08-09 04:12:17
l2_adamw2_bigtail11i_t2_sha	8fc53b3cdb1493b457b85f98797d183e9df8c93d	Thai Thien	2020-08-09 03:13:05
l2_adamw2_bigtail11i_t1_sha	5d240b5c9ea5dd91e149b1b7652e1077d1158e39	Thai Thien	2020-08-08 18:02:30
shanghaitech_keepfull_r50 l2_adamw2_bigtail13i_t6_sha	2253ce7c06882f44ed577a2df25a0b9325c476b4	Thai Thien	2020-08-08 17:57:24
l2_adamw2_bigtail13i_t5_sha	7df89278208a7d3d2117940c34c6528d38508708	Thai Thien	2020-08-08 17:42:11
lr scheduler	8de1b255ab35d2c7696fca89ea6208f6a80233da	Thai Thien	2020-08-08 16:02:42
add lr_scheduler	05704fd6f3f04bf0e41983142b3c48fec3a502ab	Thai Thien	2020-08-08 15:59:13
wd -5	768042c4ef103a94437bedca96979dce02a12c79	Thai Thien	2020-08-08 15:37:12
l2_adamw2_bigtail13i_t1_sha	3d87265cbd52127971615daf872b8467cb2baa08	Thai Thien	2020-08-08 15:02:36
let verify sha_1 to see how it is doing	88550cb75ac83c5443b0a5e5f805f0c6cc05cfde	Thai Thien	2020-08-08 09:59:21
logs with s again	8148d61da1c8eca9e4b3a4b24ddd6d44db069f0f	Thai Thien	2020-08-05 17:02:27
sh -> py	fd6e767e9ef404b41e50b7b99ca76806a65ff774	Thai Thien	2020-08-05 17:01:40
logs with "s"	59628c89e12c2c1a39f8ba97beb6cbe7e7e5c85a	Thai Thien	2020-08-05 17:00:31
verify sha	ed05894d30fa734b2d8c0ae2ee7b514018203fec	Thai Thien	2020-08-05 16:55:33

Commit 50d9652b43d660d108d697ad730e547aabf32e46 - perfomance test
Author: Thai Thien
Author date (UTC): 2020-08-14 15:46
Committer name: Thai Thien
Committer date (UTC): 2020-08-14 15:46
Parent(s): dd91bf5002f1505d1dd214c13ea416dbb610e2e3
Signer:
Signing key:
Signing status: N
Tree: e4284baa982036bfe9a687a32fe6dccecf9af9bd

File	Lines added	Lines deleted
debug/perfomance_test_on_shb.py	163	0

File debug/perfomance_test_on_shb.py added (mode: 100644) (index 0000000..cc415fd)
	1	from comet_ml import Experiment
	2
	3	from args_util import meow_parse, lr_scheduler_milestone_builder
	4	from data_flow import get_dataloader, create_image_list
	5	from ignite.engine import Events, create_supervised_trainer, create_supervised_evaluator
	6	from ignite.metrics import Loss
	7	from ignite.handlers import Checkpoint, DiskSaver, Timer
	8	from crowd_counting_error_metrics import CrowdCountingMeanAbsoluteError, CrowdCountingMeanSquaredError, CrowdCountingMeanAbsoluteErrorWithCount, CrowdCountingMeanSquaredErrorWithCount
	9	from visualize_util import get_readable_time
	10	from mse_l1_loss import MSEL1Loss, MSE4L1Loss
	11	import torch
	12	from torch import nn
	13	from models.meow_experiment.kitten_meow_1 import M1, M2, M3, M4
	14	from models.meow_experiment.ccnn_tail import BigTailM1, BigTailM2, BigTail3, BigTail4, BigTail5, BigTail6, BigTail7, BigTail8, BigTail6i, BigTail9i
	15	from models.meow_experiment.ccnn_tail import BigTail11i, BigTail10i, BigTail12i, BigTail13i, BigTail14i, BigTail15i
	16	from models.meow_experiment.ccnn_head import H1, H2, H3, H3i, H4i
	17	from models.meow_experiment.kitten_meow_1 import H1_Bigtail3
	18	from models import CustomCNNv2, CompactCNNV7
	19	from models.compact_cnn import CompactCNNV8, CompactCNNV9, CompactCNNV7i
	20	import os
	21	from model_util import get_lr, BestMetrics
	22	from ignite.contrib.handlers import PiecewiseLinear
	23	import time
	24
	25
	26
	27
	28
	29
	30
	31	"""
	32	Document on save load model
	33	https://pytorch.org/tutorials/beginner/saving_loading_models.html
	34	"""
	35
	36	model_path = "/data/save_model/adamw1_bigtail13i_t1_shb/adamw1_bigtail13i_t1_shb_checkpoint_valid_mae=-7.574910521507263.pth"
	37	checkpoint = torch.load(model_path)
	38
	39	model = BigTail13i()
	40	model.load_state_dict(checkpoint["model"])
	41	print("done load")
	42	run_test_loader(model)
	43
	44	if __name__ == "__main__":
	45	torch.set_num_threads(2) # 4 thread
	46
	47	device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
	48	print(device)
	49	args = meow_parse()
	50	print(args)
	51
	52	DATA_PATH = args.input
	53	TRAIN_PATH = os.path.join(DATA_PATH, "train_data_train_split")
	54	VAL_PATH = os.path.join(DATA_PATH, "train_data_validate_split")
	55	TEST_PATH = os.path.join(DATA_PATH, "test_data")
	56	dataset_name = args.datasetname
	57	if dataset_name=="shanghaitech":
	58	print("will use shanghaitech dataset with crop ")
	59	elif dataset_name == "shanghaitech_keepfull":
	60	print("will use shanghaitech_keepfull")
	61	else:
	62	print("cannot detect dataset_name")
	63	print("current dataset_name is ", dataset_name)
	64
	65	# create list
	66	train_list = create_image_list(TRAIN_PATH)
	67	val_list = create_image_list(VAL_PATH)
	68	test_list = create_image_list(TEST_PATH)
	69	train_loader, train_loader_eval, val_loader, test_loader = get_dataloader(train_list, val_list, test_list, dataset_name=dataset_name, batch_size=args.batch_size,
	70	train_loader_for_eval_check=True,
	71	cache=args.cache,
	72	pin_memory=args.pin_memory)
	73
	74	print("len train_loader ", len(train_loader))
	75
	76	# model
	77	model_name = args.model
	78
	79	if model_name == "M1":
	80	model = M1()
	81	elif model_name == "M2":
	82	model = M2()
	83	elif model_name == "M3":
	84	model = M3()
	85	elif model_name == "M4":
	86	model = M4()
	87	elif model_name == "CustomCNNv2":
	88	model = CustomCNNv2()
	89	elif model_name == "BigTailM1":
	90	model = BigTailM1()
	91	elif model_name == "BigTailM2":
	92	model = BigTailM2()
	93	elif model_name == "BigTail3":
	94	model = BigTail3()
	95	elif model_name == "BigTail4":
	96	model = BigTail4()
	97	elif model_name == "BigTail5":
	98	model = BigTail5()
	99	elif model_name == "BigTail6":
	100	model = BigTail6()
	101	elif model_name == "BigTail6i":
	102	model = BigTail6i()
	103	elif model_name == "BigTail9i":
	104	model = BigTail9i()
	105	elif model_name == "BigTail10i":
	106	model = BigTail10i()
	107	elif model_name == "BigTail11i":
	108	model = BigTail11i()
	109	elif model_name == "BigTail12i":
	110	model = BigTail12i()
	111	elif model_name == "BigTail13i":
	112	model = BigTail13i()
	113	elif model_name == "BigTail14i":
	114	model = BigTail14i()
	115	elif model_name == "BigTail15i":
	116	model = BigTail15i()
	117	elif model_name == "BigTail7":
	118	model = BigTail7()
	119	elif model_name == "BigTail8":
	120	model = BigTail8()
	121	elif model_name == "H1":
	122	model = H1()
	123	elif model_name == "H2":
	124	model = H2()
	125	elif model_name == "H3":
	126	model = H3()
	127	elif model_name == "H3i":
	128	model = H3i()
	129	elif model_name == "H4i":
	130	model = H4i()
	131	elif model_name == "H1_Bigtail3":
	132	model = H1_Bigtail3()
	133	elif model_name == "CompactCNNV7":
	134	model = CompactCNNV7()
	135	elif model_name == "CompactCNNV7i":
	136	model = CompactCNNV7i()
	137	elif model_name == "CompactCNNV8":
	138	model = CompactCNNV8()
	139	elif model_name == "CompactCNNV9":
	140	model = CompactCNNV9()
	141	else:
	142	print("error: you didn't pick a model")
	143	exit(-1)
	144	model = model.to(device)
	145	checkpoint = torch.load(model_path)
	146	model.load_state_dict(checkpoint["model"])
	147
	148	s1 = time.perf_counter()
	149	for img, label in test_loader:
	150	pred = model(img.cuda())
	151	print("done")
	152	s2 = time.perf_counter()
	153	time1 = s1 - s2
	154	print("test 1 time " + str(s1 - s2))
	155
	156	s3 = time.perf_counter()
	157	for img, label in test_loader:
	158	pred = model(img.cuda())
	159	print("done")
	160	s4 = time.perf_counter()
	161	time1 = s3 - s4
	162	print("test 1 time " + str(s3 - s4))
	163

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