Commit
•
e88e8f7
1
Parent(s):
2683b18
W&B: Restructure code to support the new dataset_check() feature (#4197)
Browse files* Improve docstrings and run names
* default wandb login prompt with timeout
* return key
* Update api_key check logic
* Properly support zipped dataset feature
* update docstring
* Revert tuorial change
* extend changes to log_dataset
* add run name
* bug fix
* bug fix
* Update comment
* fix import check
* remove unused import
* Hardcore .yaml file extension
* reduce code
* Reformat using pycharm
Co-authored-by: Glenn Jocher <glenn.jocher@ultralytics.com>
- README.md +0 -0
- train.py +11 -6
- utils/loggers/__init__.py +4 -9
- utils/loggers/wandb/log_dataset.py +2 -4
- utils/loggers/wandb/sweep.py +1 -2
- utils/loggers/wandb/wandb_utils.py +34 -19
README.md
CHANGED
File without changes
|
train.py
CHANGED
@@ -73,24 +73,29 @@ def train(hyp, # path/to/hyp.yaml or hyp dictionary
|
|
73 |
yaml.safe_dump(hyp, f, sort_keys=False)
|
74 |
with open(save_dir / 'opt.yaml', 'w') as f:
|
75 |
yaml.safe_dump(vars(opt), f, sort_keys=False)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
76 |
|
|
|
77 |
# Config
|
78 |
plots = not evolve # create plots
|
79 |
cuda = device.type != 'cpu'
|
80 |
init_seeds(1 + RANK)
|
81 |
with torch_distributed_zero_first(RANK):
|
82 |
-
data_dict = check_dataset(data) # check
|
83 |
train_path, val_path = data_dict['train'], data_dict['val']
|
84 |
nc = 1 if single_cls else int(data_dict['nc']) # number of classes
|
85 |
names = ['item'] if single_cls and len(data_dict['names']) != 1 else data_dict['names'] # class names
|
86 |
assert len(names) == nc, f'{len(names)} names found for nc={nc} dataset in {data}' # check
|
87 |
is_coco = data.endswith('coco.yaml') and nc == 80 # COCO dataset
|
88 |
|
89 |
-
# Loggers
|
90 |
-
if RANK in [-1, 0]:
|
91 |
-
loggers = Loggers(save_dir, weights, opt, hyp, data_dict, LOGGER).start() # loggers dict
|
92 |
-
if loggers.wandb and resume:
|
93 |
-
weights, epochs, hyp, data_dict = opt.weights, opt.epochs, opt.hyp, loggers.wandb.data_dict
|
94 |
|
95 |
# Model
|
96 |
pretrained = weights.endswith('.pt')
|
|
|
73 |
yaml.safe_dump(hyp, f, sort_keys=False)
|
74 |
with open(save_dir / 'opt.yaml', 'w') as f:
|
75 |
yaml.safe_dump(vars(opt), f, sort_keys=False)
|
76 |
+
data_dict = None
|
77 |
+
|
78 |
+
# Loggers
|
79 |
+
if RANK in [-1, 0]:
|
80 |
+
loggers = Loggers(save_dir, weights, opt, hyp, LOGGER).start() # loggers dict
|
81 |
+
if loggers.wandb:
|
82 |
+
data_dict = loggers.wandb.data_dict
|
83 |
+
if resume:
|
84 |
+
weights, epochs, hyp = opt.weights, opt.epochs, opt.hyp
|
85 |
|
86 |
+
|
87 |
# Config
|
88 |
plots = not evolve # create plots
|
89 |
cuda = device.type != 'cpu'
|
90 |
init_seeds(1 + RANK)
|
91 |
with torch_distributed_zero_first(RANK):
|
92 |
+
data_dict = data_dict or check_dataset(data) # check if None
|
93 |
train_path, val_path = data_dict['train'], data_dict['val']
|
94 |
nc = 1 if single_cls else int(data_dict['nc']) # number of classes
|
95 |
names = ['item'] if single_cls and len(data_dict['names']) != 1 else data_dict['names'] # class names
|
96 |
assert len(names) == nc, f'{len(names)} names found for nc={nc} dataset in {data}' # check
|
97 |
is_coco = data.endswith('coco.yaml') and nc == 80 # COCO dataset
|
98 |
|
|
|
|
|
|
|
|
|
|
|
99 |
|
100 |
# Model
|
101 |
pretrained = weights.endswith('.pt')
|
utils/loggers/__init__.py
CHANGED
@@ -1,9 +1,7 @@
|
|
1 |
# YOLOv5 experiment logging utils
|
2 |
-
|
3 |
import warnings
|
4 |
from threading import Thread
|
5 |
-
|
6 |
-
import torch
|
7 |
from torch.utils.tensorboard import SummaryWriter
|
8 |
|
9 |
from utils.general import colorstr, emojis
|
@@ -23,12 +21,11 @@ except (ImportError, AssertionError):
|
|
23 |
|
24 |
class Loggers():
|
25 |
# YOLOv5 Loggers class
|
26 |
-
def __init__(self, save_dir=None, weights=None, opt=None, hyp=None,
|
27 |
self.save_dir = save_dir
|
28 |
self.weights = weights
|
29 |
self.opt = opt
|
30 |
self.hyp = hyp
|
31 |
-
self.data_dict = data_dict
|
32 |
self.logger = logger # for printing results to console
|
33 |
self.include = include
|
34 |
for k in LOGGERS:
|
@@ -38,9 +35,7 @@ class Loggers():
|
|
38 |
self.csv = True # always log to csv
|
39 |
|
40 |
# Message
|
41 |
-
|
42 |
-
import wandb
|
43 |
-
except ImportError:
|
44 |
prefix = colorstr('Weights & Biases: ')
|
45 |
s = f"{prefix}run 'pip install wandb' to automatically track and visualize YOLOv5 🚀 runs (RECOMMENDED)"
|
46 |
print(emojis(s))
|
@@ -57,7 +52,7 @@ class Loggers():
|
|
57 |
assert 'wandb' in self.include and wandb
|
58 |
run_id = torch.load(self.weights).get('wandb_id') if self.opt.resume else None
|
59 |
self.opt.hyp = self.hyp # add hyperparameters
|
60 |
-
self.wandb = WandbLogger(self.opt, run_id
|
61 |
except:
|
62 |
self.wandb = None
|
63 |
|
|
|
1 |
# YOLOv5 experiment logging utils
|
2 |
+
import torch
|
3 |
import warnings
|
4 |
from threading import Thread
|
|
|
|
|
5 |
from torch.utils.tensorboard import SummaryWriter
|
6 |
|
7 |
from utils.general import colorstr, emojis
|
|
|
21 |
|
22 |
class Loggers():
|
23 |
# YOLOv5 Loggers class
|
24 |
+
def __init__(self, save_dir=None, weights=None, opt=None, hyp=None, logger=None, include=LOGGERS):
|
25 |
self.save_dir = save_dir
|
26 |
self.weights = weights
|
27 |
self.opt = opt
|
28 |
self.hyp = hyp
|
|
|
29 |
self.logger = logger # for printing results to console
|
30 |
self.include = include
|
31 |
for k in LOGGERS:
|
|
|
35 |
self.csv = True # always log to csv
|
36 |
|
37 |
# Message
|
38 |
+
if not wandb:
|
|
|
|
|
39 |
prefix = colorstr('Weights & Biases: ')
|
40 |
s = f"{prefix}run 'pip install wandb' to automatically track and visualize YOLOv5 🚀 runs (RECOMMENDED)"
|
41 |
print(emojis(s))
|
|
|
52 |
assert 'wandb' in self.include and wandb
|
53 |
run_id = torch.load(self.weights).get('wandb_id') if self.opt.resume else None
|
54 |
self.opt.hyp = self.hyp # add hyperparameters
|
55 |
+
self.wandb = WandbLogger(self.opt, run_id)
|
56 |
except:
|
57 |
self.wandb = None
|
58 |
|
utils/loggers/wandb/log_dataset.py
CHANGED
@@ -1,5 +1,4 @@
|
|
1 |
import argparse
|
2 |
-
|
3 |
import yaml
|
4 |
|
5 |
from wandb_utils import WandbLogger
|
@@ -8,9 +7,7 @@ WANDB_ARTIFACT_PREFIX = 'wandb-artifact://'
|
|
8 |
|
9 |
|
10 |
def create_dataset_artifact(opt):
|
11 |
-
|
12 |
-
data = yaml.safe_load(f) # data dict
|
13 |
-
logger = WandbLogger(opt, '', None, data, job_type='Dataset Creation') # TODO: return value unused
|
14 |
|
15 |
|
16 |
if __name__ == '__main__':
|
@@ -19,6 +16,7 @@ if __name__ == '__main__':
|
|
19 |
parser.add_argument('--single-cls', action='store_true', help='train as single-class dataset')
|
20 |
parser.add_argument('--project', type=str, default='YOLOv5', help='name of W&B Project')
|
21 |
parser.add_argument('--entity', default=None, help='W&B entity')
|
|
|
22 |
|
23 |
opt = parser.parse_args()
|
24 |
opt.resume = False # Explicitly disallow resume check for dataset upload job
|
|
|
1 |
import argparse
|
|
|
2 |
import yaml
|
3 |
|
4 |
from wandb_utils import WandbLogger
|
|
|
7 |
|
8 |
|
9 |
def create_dataset_artifact(opt):
|
10 |
+
logger = WandbLogger(opt, None, job_type='Dataset Creation') # TODO: return value unused
|
|
|
|
|
11 |
|
12 |
|
13 |
if __name__ == '__main__':
|
|
|
16 |
parser.add_argument('--single-cls', action='store_true', help='train as single-class dataset')
|
17 |
parser.add_argument('--project', type=str, default='YOLOv5', help='name of W&B Project')
|
18 |
parser.add_argument('--entity', default=None, help='W&B entity')
|
19 |
+
parser.add_argument('--name', type=str, default='log dataset', help='name of W&B run')
|
20 |
|
21 |
opt = parser.parse_args()
|
22 |
opt.resume = False # Explicitly disallow resume check for dataset upload job
|
utils/loggers/wandb/sweep.py
CHANGED
@@ -1,7 +1,6 @@
|
|
1 |
import sys
|
2 |
-
from pathlib import Path
|
3 |
-
|
4 |
import wandb
|
|
|
5 |
|
6 |
FILE = Path(__file__).absolute()
|
7 |
sys.path.append(FILE.parents[2].as_posix()) # add utils/ to path
|
|
|
1 |
import sys
|
|
|
|
|
2 |
import wandb
|
3 |
+
from pathlib import Path
|
4 |
|
5 |
FILE = Path(__file__).absolute()
|
6 |
sys.path.append(FILE.parents[2].as_posix()) # add utils/ to path
|
utils/loggers/wandb/wandb_utils.py
CHANGED
@@ -3,10 +3,9 @@
|
|
3 |
import logging
|
4 |
import os
|
5 |
import sys
|
|
|
6 |
from contextlib import contextmanager
|
7 |
from pathlib import Path
|
8 |
-
|
9 |
-
import yaml
|
10 |
from tqdm import tqdm
|
11 |
|
12 |
FILE = Path(__file__).absolute()
|
@@ -99,7 +98,7 @@ class WandbLogger():
|
|
99 |
https://docs.wandb.com/guides/integrations/yolov5
|
100 |
"""
|
101 |
|
102 |
-
def __init__(self, opt, run_id,
|
103 |
"""
|
104 |
- Initialize WandbLogger instance
|
105 |
- Upload dataset if opt.upload_dataset is True
|
@@ -108,7 +107,6 @@ class WandbLogger():
|
|
108 |
arguments:
|
109 |
opt (namespace) -- Commandline arguments for this run
|
110 |
run_id (str) -- Run ID of W&B run to be resumed
|
111 |
-
data_dict (Dict) -- Dictionary conataining info about the dataset to be used
|
112 |
job_type (str) -- To set the job_type for this run
|
113 |
|
114 |
"""
|
@@ -119,10 +117,11 @@ class WandbLogger():
|
|
119 |
self.train_artifact_path, self.val_artifact_path = None, None
|
120 |
self.result_artifact = None
|
121 |
self.val_table, self.result_table = None, None
|
122 |
-
self.data_dict = data_dict
|
123 |
self.bbox_media_panel_images = []
|
124 |
self.val_table_path_map = None
|
125 |
self.max_imgs_to_log = 16
|
|
|
|
|
126 |
# It's more elegant to stick to 1 wandb.init call, but useful config data is overwritten in the WandbLogger's wandb.init call
|
127 |
if isinstance(opt.resume, str): # checks resume from artifact
|
128 |
if opt.resume.startswith(WANDB_ARTIFACT_PREFIX):
|
@@ -148,11 +147,23 @@ class WandbLogger():
|
|
148 |
if self.wandb_run:
|
149 |
if self.job_type == 'Training':
|
150 |
if not opt.resume:
|
151 |
-
|
152 |
-
|
153 |
-
|
154 |
-
|
155 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
156 |
if self.job_type == 'Dataset Creation':
|
157 |
self.data_dict = self.check_and_upload_dataset(opt)
|
158 |
|
@@ -167,7 +178,7 @@ class WandbLogger():
|
|
167 |
Updated dataset info dictionary where local dataset paths are replaced by WAND_ARFACT_PREFIX links.
|
168 |
"""
|
169 |
assert wandb, 'Install wandb to upload dataset'
|
170 |
-
config_path = self.log_dataset_artifact(
|
171 |
opt.single_cls,
|
172 |
'YOLOv5' if opt.project == 'runs/train' else Path(opt.project).stem)
|
173 |
print("Created dataset config file ", config_path)
|
@@ -175,7 +186,7 @@ class WandbLogger():
|
|
175 |
wandb_data_dict = yaml.safe_load(f)
|
176 |
return wandb_data_dict
|
177 |
|
178 |
-
def setup_training(self, opt
|
179 |
"""
|
180 |
Setup the necessary processes for training YOLO models:
|
181 |
- Attempt to download model checkpoint and dataset artifacts if opt.resume stats with WANDB_ARTIFACT_PREFIX
|
@@ -184,10 +195,7 @@ class WandbLogger():
|
|
184 |
|
185 |
arguments:
|
186 |
opt (namespace) -- commandline arguments for this run
|
187 |
-
data_dict (Dict) -- Dataset dictionary for this run
|
188 |
|
189 |
-
returns:
|
190 |
-
data_dict (Dict) -- contains the updated info about the dataset to be used for training
|
191 |
"""
|
192 |
self.log_dict, self.current_epoch = {}, 0
|
193 |
self.bbox_interval = opt.bbox_interval
|
@@ -198,8 +206,10 @@ class WandbLogger():
|
|
198 |
config = self.wandb_run.config
|
199 |
opt.weights, opt.save_period, opt.batch_size, opt.bbox_interval, opt.epochs, opt.hyp = str(
|
200 |
self.weights), config.save_period, config.batch_size, config.bbox_interval, config.epochs, \
|
201 |
-
config.
|
202 |
data_dict = dict(self.wandb_run.config.data_dict) # eliminates the need for config file to resume
|
|
|
|
|
203 |
if self.val_artifact is None: # If --upload_dataset is set, use the existing artifact, don't download
|
204 |
self.train_artifact_path, self.train_artifact = self.download_dataset_artifact(data_dict.get('train'),
|
205 |
opt.artifact_alias)
|
@@ -221,7 +231,10 @@ class WandbLogger():
|
|
221 |
self.map_val_table_path()
|
222 |
if opt.bbox_interval == -1:
|
223 |
self.bbox_interval = opt.bbox_interval = (opt.epochs // 10) if opt.epochs > 10 else 1
|
224 |
-
|
|
|
|
|
|
|
225 |
|
226 |
def download_dataset_artifact(self, path, alias):
|
227 |
"""
|
@@ -299,7 +312,8 @@ class WandbLogger():
|
|
299 |
returns:
|
300 |
the new .yaml file with artifact links. it can be used to start training directly from artifacts
|
301 |
"""
|
302 |
-
|
|
|
303 |
nc, names = (1, ['item']) if single_cls else (int(data['nc']), data['names'])
|
304 |
names = {k: v for k, v in enumerate(names)} # to index dictionary
|
305 |
self.train_artifact = self.create_dataset_table(LoadImagesAndLabels(
|
@@ -310,7 +324,8 @@ class WandbLogger():
|
|
310 |
data['train'] = WANDB_ARTIFACT_PREFIX + str(Path(project) / 'train')
|
311 |
if data.get('val'):
|
312 |
data['val'] = WANDB_ARTIFACT_PREFIX + str(Path(project) / 'val')
|
313 |
-
path =
|
|
|
314 |
data.pop('download', None)
|
315 |
data.pop('path', None)
|
316 |
with open(path, 'w') as f:
|
|
|
3 |
import logging
|
4 |
import os
|
5 |
import sys
|
6 |
+
import yaml
|
7 |
from contextlib import contextmanager
|
8 |
from pathlib import Path
|
|
|
|
|
9 |
from tqdm import tqdm
|
10 |
|
11 |
FILE = Path(__file__).absolute()
|
|
|
98 |
https://docs.wandb.com/guides/integrations/yolov5
|
99 |
"""
|
100 |
|
101 |
+
def __init__(self, opt, run_id, job_type='Training'):
|
102 |
"""
|
103 |
- Initialize WandbLogger instance
|
104 |
- Upload dataset if opt.upload_dataset is True
|
|
|
107 |
arguments:
|
108 |
opt (namespace) -- Commandline arguments for this run
|
109 |
run_id (str) -- Run ID of W&B run to be resumed
|
|
|
110 |
job_type (str) -- To set the job_type for this run
|
111 |
|
112 |
"""
|
|
|
117 |
self.train_artifact_path, self.val_artifact_path = None, None
|
118 |
self.result_artifact = None
|
119 |
self.val_table, self.result_table = None, None
|
|
|
120 |
self.bbox_media_panel_images = []
|
121 |
self.val_table_path_map = None
|
122 |
self.max_imgs_to_log = 16
|
123 |
+
self.wandb_artifact_data_dict = None
|
124 |
+
self.data_dict = None
|
125 |
# It's more elegant to stick to 1 wandb.init call, but useful config data is overwritten in the WandbLogger's wandb.init call
|
126 |
if isinstance(opt.resume, str): # checks resume from artifact
|
127 |
if opt.resume.startswith(WANDB_ARTIFACT_PREFIX):
|
|
|
147 |
if self.wandb_run:
|
148 |
if self.job_type == 'Training':
|
149 |
if not opt.resume:
|
150 |
+
if opt.upload_dataset:
|
151 |
+
self.wandb_artifact_data_dict = self.check_and_upload_dataset(opt)
|
152 |
+
|
153 |
+
elif opt.data.endswith('_wandb.yaml'): # When dataset is W&B artifact
|
154 |
+
with open(opt.data, encoding='ascii', errors='ignore') as f:
|
155 |
+
data_dict = yaml.safe_load(f)
|
156 |
+
self.data_dict = data_dict
|
157 |
+
else: # Local .yaml dataset file or .zip file
|
158 |
+
self.data_dict = check_dataset(opt.data)
|
159 |
+
|
160 |
+
self.setup_training(opt)
|
161 |
+
# write data_dict to config. useful for resuming from artifacts
|
162 |
+
if not self.wandb_artifact_data_dict:
|
163 |
+
self.wandb_artifact_data_dict = self.data_dict
|
164 |
+
self.wandb_run.config.update({'data_dict': self.wandb_artifact_data_dict},
|
165 |
+
allow_val_change=True)
|
166 |
+
|
167 |
if self.job_type == 'Dataset Creation':
|
168 |
self.data_dict = self.check_and_upload_dataset(opt)
|
169 |
|
|
|
178 |
Updated dataset info dictionary where local dataset paths are replaced by WAND_ARFACT_PREFIX links.
|
179 |
"""
|
180 |
assert wandb, 'Install wandb to upload dataset'
|
181 |
+
config_path = self.log_dataset_artifact(opt.data,
|
182 |
opt.single_cls,
|
183 |
'YOLOv5' if opt.project == 'runs/train' else Path(opt.project).stem)
|
184 |
print("Created dataset config file ", config_path)
|
|
|
186 |
wandb_data_dict = yaml.safe_load(f)
|
187 |
return wandb_data_dict
|
188 |
|
189 |
+
def setup_training(self, opt):
|
190 |
"""
|
191 |
Setup the necessary processes for training YOLO models:
|
192 |
- Attempt to download model checkpoint and dataset artifacts if opt.resume stats with WANDB_ARTIFACT_PREFIX
|
|
|
195 |
|
196 |
arguments:
|
197 |
opt (namespace) -- commandline arguments for this run
|
|
|
198 |
|
|
|
|
|
199 |
"""
|
200 |
self.log_dict, self.current_epoch = {}, 0
|
201 |
self.bbox_interval = opt.bbox_interval
|
|
|
206 |
config = self.wandb_run.config
|
207 |
opt.weights, opt.save_period, opt.batch_size, opt.bbox_interval, opt.epochs, opt.hyp = str(
|
208 |
self.weights), config.save_period, config.batch_size, config.bbox_interval, config.epochs, \
|
209 |
+
config.hyp
|
210 |
data_dict = dict(self.wandb_run.config.data_dict) # eliminates the need for config file to resume
|
211 |
+
else:
|
212 |
+
data_dict = self.data_dict
|
213 |
if self.val_artifact is None: # If --upload_dataset is set, use the existing artifact, don't download
|
214 |
self.train_artifact_path, self.train_artifact = self.download_dataset_artifact(data_dict.get('train'),
|
215 |
opt.artifact_alias)
|
|
|
231 |
self.map_val_table_path()
|
232 |
if opt.bbox_interval == -1:
|
233 |
self.bbox_interval = opt.bbox_interval = (opt.epochs // 10) if opt.epochs > 10 else 1
|
234 |
+
train_from_artifact = self.train_artifact_path is not None and self.val_artifact_path is not None
|
235 |
+
# Update the the data_dict to point to local artifacts dir
|
236 |
+
if train_from_artifact:
|
237 |
+
self.data_dict = data_dict
|
238 |
|
239 |
def download_dataset_artifact(self, path, alias):
|
240 |
"""
|
|
|
312 |
returns:
|
313 |
the new .yaml file with artifact links. it can be used to start training directly from artifacts
|
314 |
"""
|
315 |
+
self.data_dict = check_dataset(data_file) # parse and check
|
316 |
+
data = dict(self.data_dict)
|
317 |
nc, names = (1, ['item']) if single_cls else (int(data['nc']), data['names'])
|
318 |
names = {k: v for k, v in enumerate(names)} # to index dictionary
|
319 |
self.train_artifact = self.create_dataset_table(LoadImagesAndLabels(
|
|
|
324 |
data['train'] = WANDB_ARTIFACT_PREFIX + str(Path(project) / 'train')
|
325 |
if data.get('val'):
|
326 |
data['val'] = WANDB_ARTIFACT_PREFIX + str(Path(project) / 'val')
|
327 |
+
path = Path(data_file).stem
|
328 |
+
path = (path if overwrite_config else path + '_wandb') + '.yaml' # updated data.yaml path
|
329 |
data.pop('download', None)
|
330 |
data.pop('path', None)
|
331 |
with open(path, 'w') as f:
|