HongChuang/ultralytics-ascend
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 3

ultralytics 8.0.174 Tuner plots and improvements (#4799)

Browse source 
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
This commit is contained in:
Glenn Jocher 2023-09-10 03:27:23 +02:00 • committed by GitHub
parent dfe6dfb1d2
commit 16ce193d6e
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
12 changed files with 248 additions and 88 deletions
Download patch file Download diff file Expand all files Collapse all files

111
ultralytics/engine/tuner.py
View file

@ -13,18 +13,20 @@ Example:
from ultralytics import YOLO
model = YOLO('yolov8n.pt')
model.tune(data='coco8.yaml', imgsz=640, epochs=100, iterations=10)
model.tune(data='coco8.yaml', epochs=10, iterations=300, optimizer='AdamW', plots=False, save=False, val=False)
```
"""
import random
import shutil
import subprocess
import time
from copy import deepcopy
import numpy as np
import torch
from ultralytics import YOLO
from ultralytics.cfg import get_cfg, get_save_dir
from ultralytics.utils import DEFAULT_CFG, LOGGER, callbacks, colorstr, yaml_print, yaml_save
from ultralytics.utils import DEFAULT_CFG, LOGGER, callbacks, colorstr, remove_colorstr, yaml_print, yaml_save
from ultralytics.utils.plotting import plot_tune_results
class Tuner:
@ -37,7 +39,7 @@ class Tuner:
Attributes:
space (dict): Hyperparameter search space containing bounds and scaling factors for mutation.
tune_dir (Path): Directory where evolution logs and results will be saved.
evolve_csv (Path): Path to the CSV file where evolution logs are saved.
tune_csv (Path): Path to the CSV file where evolution logs are saved.
Methods:
_mutate(hyp: dict) -> dict:
@ -52,7 +54,7 @@ class Tuner:
from ultralytics import YOLO
model = YOLO('yolov8n.pt')
model.tune(data='coco8.yaml', imgsz=640, epochs=100, iterations=10, val=False, cache=True)
model.tune(data='coco8.yaml', epochs=10, iterations=300, optimizer='AdamW', plots=False, save=False, val=False)
```
"""
@ -64,22 +66,23 @@ class Tuner:
args (dict, optional): Configuration for hyperparameter evolution.
"""
self.args = get_cfg(overrides=args)
self.space = { # key: (min, max, gain(optionaL))
self.space = { # key: (min, max, gain(optional))
# 'optimizer': tune.choice(['SGD', 'Adam', 'AdamW', 'NAdam', 'RAdam', 'RMSProp']),
'lr0': (1e-5, 1e-1),
'lrf': (0.01, 1.0), # final OneCycleLR learning rate (lr0 * lrf)
'momentum': (0.6, 0.98, 0.3), # SGD momentum/Adam beta1
'lrf': (0.0001, 0.1), # final OneCycleLR learning rate (lr0 * lrf)
'momentum': (0.7, 0.98, 0.3), # SGD momentum/Adam beta1
'weight_decay': (0.0, 0.001), # optimizer weight decay 5e-4
'warmup_epochs': (0.0, 5.0), # warmup epochs (fractions ok)
'warmup_momentum': (0.0, 0.95), # warmup initial momentum
'box': (0.02, 0.2), # box loss gain
'box': (1.0, 20.0), # box loss gain
'cls': (0.2, 4.0), # cls loss gain (scale with pixels)
'dfl': (0.4, 6.0), # dfl loss gain
'hsv_h': (0.0, 0.1), # image HSV-Hue augmentation (fraction)
'hsv_s': (0.0, 0.9), # image HSV-Saturation augmentation (fraction)
'hsv_v': (0.0, 0.9), # image HSV-Value augmentation (fraction)
'degrees': (0.0, 45.0), # image rotation (+/- deg)
'translate': (0.0, 0.9), # image translation (+/- fraction)
'scale': (0.0, 0.9), # image scale (+/- gain)
'scale': (0.0, 0.95), # image scale (+/- gain)
'shear': (0.0, 10.0), # image shear (+/- deg)
'perspective': (0.0, 0.001), # image perspective (+/- fraction), range 0-0.001
'flipud': (0.0, 1.0), # image flip up-down (probability)
@ -87,11 +90,13 @@ class Tuner:
'mosaic': (0.0, 1.0), # image mixup (probability)
'mixup': (0.0, 1.0), # image mixup (probability)
'copy_paste': (0.0, 1.0)} # segment copy-paste (probability)
self.tune_dir = get_save_dir(self.args, name='_tune')
self.evolve_csv = self.tune_dir / 'evolve.csv'
self.tune_dir = get_save_dir(self.args, name='tune')
self.tune_csv = self.tune_dir / 'tune_results.csv'
self.callbacks = _callbacks or callbacks.get_default_callbacks()
self.prefix = colorstr('Tuner: ')
callbacks.add_integration_callbacks(self)
LOGGER.info(f"Initialized Tuner instance with 'tune_dir={self.tune_dir}'.")
LOGGER.info(f"{self.prefix}Initialized Tuner instance with 'tune_dir={self.tune_dir}'\n"
f'{self.prefix}💡 Learn about tuning at https://docs.ultralytics.com/guides/hyperparameter-tuning')
def _mutate(self, parent='single', n=5, mutation=0.8, sigma=0.2):
"""
@ -106,9 +111,9 @@ class Tuner:
Returns:
(dict): A dictionary containing mutated hyperparameters.
"""
if self.evolve_csv.exists(): # if evolve.csv exists: select best hyps and mutate
if self.tune_csv.exists(): # if CSV file exists: select best hyps and mutate
# Select parent(s)
x = np.loadtxt(self.evolve_csv, ndmin=2, delimiter=',', skiprows=1)
x = np.loadtxt(self.tune_csv, ndmin=2, delimiter=',', skiprows=1)
fitness = x[:, 0] # first column
n = min(n, len(x)) # number of previous results to consider
x = x[np.argsort(-fitness)][:n] # top n mutations
@ -139,7 +144,7 @@ class Tuner:
return hyp
def __call__(self, model=None, iterations=10, prefix=colorstr('Tuner:')):
def __call__(self, model=None, iterations=10, cleanup=True):
"""
Executes the hyperparameter evolution process when the Tuner instance is called.
@ -152,54 +157,68 @@ class Tuner:
Args:
model (Model): A pre-initialized YOLO model to be used for training.
iterations (int): The number of generations to run the evolution for.
cleanup (bool): Whether to delete iteration weights to reduce storage space used during tuning.
Note:
The method utilizes the `self.evolve_csv` Path object to read and log hyperparameters and fitness scores.
The method utilizes the `self.tune_csv` Path object to read and log hyperparameters and fitness scores.
Ensure this path is set correctly in the Tuner instance.
"""
t0 = time.time()
best_save_dir, best_metrics = None, None
self.tune_dir.mkdir(parents=True, exist_ok=True)
(self.tune_dir / 'weights').mkdir(parents=True, exist_ok=True)
for i in range(iterations):
# Mutate hyperparameters
mutated_hyp = self._mutate()
LOGGER.info(f'{prefix} Starting iteration {i + 1}/{iterations} with hyperparameters: {mutated_hyp}')
LOGGER.info(f'{self.prefix}Starting iteration {i + 1}/{iterations} with hyperparameters: {mutated_hyp}')
metrics = {}
train_args = {**vars(self.args), **mutated_hyp}
save_dir = get_save_dir(get_cfg(train_args))
try:
# Train YOLO model with mutated hyperparameters
train_args = {**vars(self.args), **mutated_hyp}
results = (deepcopy(model) or YOLO(self.args.model)).train(**train_args)
fitness = results.fitness
except Exception as e:
LOGGER.warning(f'WARNING ❌️ training failure for hyperparameter tuning iteration {i}\n{e}')
fitness = 0.0
# Train YOLO model with mutated hyperparameters (run in subprocess to avoid dataloader hang)
weights_dir = save_dir / 'weights'
cmd = ['yolo', 'train', *(f'{k}={v}' for k, v in train_args.items())]
assert subprocess.run(cmd, check=True).returncode == 0, 'training failed'
ckpt_file = weights_dir / ('best.pt' if (weights_dir / 'best.pt').exists() else 'last.pt')
metrics = torch.load(ckpt_file)['train_metrics']
# Save results and mutated_hyp to evolve_csv
except Exception as e:
LOGGER.warning(f'WARNING ❌️ training failure for hyperparameter tuning iteration {i + 1}\n{e}')
# Save results and mutated_hyp to CSV
fitness = metrics.get('fitness', 0.0)
log_row = [round(fitness, 5)] + [mutated_hyp[k] for k in self.space.keys()]
headers = '' if self.evolve_csv.exists() else (','.join(['fitness_score'] + list(self.space.keys())) + '\n')
with open(self.evolve_csv, 'a') as f:
headers = '' if self.tune_csv.exists() else (','.join(['fitness'] + list(self.space.keys())) + '\n')
with open(self.tune_csv, 'a') as f:
f.write(headers + ','.join(map(str, log_row)) + '\n')
# Print tuning results
x = np.loadtxt(self.evolve_csv, ndmin=2, delimiter=',', skiprows=1)
# Get best results
x = np.loadtxt(self.tune_csv, ndmin=2, delimiter=',', skiprows=1)
fitness = x[:, 0] # first column
best_idx = fitness.argmax()
best_is_current = best_idx == i
if best_is_current:
best_save_dir = results.save_dir
best_metrics = {k: round(v, 5) for k, v in results.results_dict.items()}
header = (f'{prefix} {i + 1} iterations complete ✅ ({time.time() - t0:.2f}s)\n'
f'{prefix} Results saved to {colorstr("bold", self.tune_dir)}\n'
f'{prefix} Best fitness={fitness[best_idx]} observed at iteration {best_idx + 1}\n'
f'{prefix} Best fitness metrics are {best_metrics}\n'
f'{prefix} Best fitness model is {best_save_dir}\n'
f'{prefix} Best fitness hyperparameters are printed below.\n')
best_save_dir = save_dir
best_metrics = {k: round(v, 5) for k, v in metrics.items()}
for ckpt in weights_dir.glob('*.pt'):
shutil.copy2(ckpt, self.tune_dir / 'weights')
elif cleanup:
shutil.rmtree(ckpt_file.parent) # remove iteration weights/ dir to reduce storage space
# Plot tune results
plot_tune_results(self.tune_csv)
# Save and print tune results
header = (f'{self.prefix}{i + 1}/{iterations} iterations complete ✅ ({time.time() - t0:.2f}s)\n'
f'{self.prefix}Results saved to {colorstr("bold", self.tune_dir)}\n'
f'{self.prefix}Best fitness={fitness[best_idx]} observed at iteration {best_idx + 1}\n'
f'{self.prefix}Best fitness metrics are {best_metrics}\n'
f'{self.prefix}Best fitness model is {best_save_dir}\n'
f'{self.prefix}Best fitness hyperparameters are printed below.\n')
LOGGER.info('\n' + header)
# Save turning results
data = {k: float(x[0, i + 1]) for i, k in enumerate(self.space.keys())}
header = header.replace(prefix, '#').replace('/', '').replace('', '') + '\n'
yaml_save(self.tune_dir / 'best.yaml', data=data, header=header)
yaml_print(self.tune_dir / 'best.yaml')
data = {k: float(x[best_idx, i + 1]) for i, k in enumerate(self.space.keys())}
yaml_save(self.tune_dir / 'best_hyperparameters.yaml',
data=data,
header=remove_colorstr(header.replace(self.prefix, '# ')) + '\n')
yaml_print(self.tune_dir / 'best_hyperparameters.yaml')