mirror of
https://github.com/open-mmlab/mmselfsup.git
synced 2025-06-03 14:59:38 +08:00
af331b043f
* [Refactor] change imgs_per_gpu to samples_per_gpu in config files * [Docs] change imgs_per_gpu to samples_per_gpu in docs * [Refactor] change imgs_per_gpu to samples_per_gpu in codes and add warnings * [Fix] fix isort * [Docs] fix docs format * [Refactor] add related UT codes * [Fix] fix isort
77 lines
2.3 KiB
Python
77 lines
2.3 KiB
Python
_base_ = [
|
|
'../_base_/models/mocov3_vit-small-p16.py',
|
|
'../_base_/datasets/imagenet_mocov3.py',
|
|
'../_base_/schedules/adamw_coslr-300e_in1k.py',
|
|
'../_base_/default_runtime.py',
|
|
]
|
|
|
|
# dataset settings
|
|
img_norm_cfg = dict(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
|
|
# the difference between ResNet50 and ViT pipeline is the `scale` in
|
|
# `RandomResizedCrop`, `scale=(0.08, 1.)` in ViT pipeline
|
|
train_pipeline1 = [
|
|
dict(type='RandomResizedCrop', size=224, scale=(0.08, 1.)),
|
|
dict(
|
|
type='RandomAppliedTrans',
|
|
transforms=[
|
|
dict(
|
|
type='ColorJitter',
|
|
brightness=0.4,
|
|
contrast=0.4,
|
|
saturation=0.2,
|
|
hue=0.1)
|
|
],
|
|
p=0.8),
|
|
dict(type='RandomGrayscale', p=0.2),
|
|
dict(type='GaussianBlur', sigma_min=0.1, sigma_max=2.0, p=1.),
|
|
dict(type='Solarization', p=0.),
|
|
dict(type='RandomHorizontalFlip'),
|
|
]
|
|
train_pipeline2 = [
|
|
dict(type='RandomResizedCrop', size=224, scale=(0.08, 1.)),
|
|
dict(
|
|
type='RandomAppliedTrans',
|
|
transforms=[
|
|
dict(
|
|
type='ColorJitter',
|
|
brightness=0.4,
|
|
contrast=0.4,
|
|
saturation=0.2,
|
|
hue=0.1)
|
|
],
|
|
p=0.8),
|
|
dict(type='RandomGrayscale', p=0.2),
|
|
dict(type='GaussianBlur', sigma_min=0.1, sigma_max=2.0, p=0.1),
|
|
dict(type='Solarization', p=0.2),
|
|
dict(type='RandomHorizontalFlip'),
|
|
]
|
|
|
|
# prefetch
|
|
prefetch = False
|
|
if not prefetch:
|
|
train_pipeline1.extend(
|
|
[dict(type='ToTensor'),
|
|
dict(type='Normalize', **img_norm_cfg)])
|
|
train_pipeline2.extend(
|
|
[dict(type='ToTensor'),
|
|
dict(type='Normalize', **img_norm_cfg)])
|
|
|
|
# dataset summary
|
|
data = dict(
|
|
samples_per_gpu=128,
|
|
train=dict(pipelines=[train_pipeline1, train_pipeline2]))
|
|
|
|
# MoCo v3 use the same momentum update method as BYOL
|
|
custom_hooks = [dict(type='MomentumUpdateHook')]
|
|
|
|
# optimizer
|
|
optimizer = dict(type='AdamW', lr=2.4e-3, weight_decay=0.1)
|
|
|
|
# fp16
|
|
fp16 = dict(loss_scale='dynamic')
|
|
|
|
# the max_keep_ckpts controls the max number of ckpt file in your work_dirs
|
|
# if it is 3, when CheckpointHook (in mmcv) saves the 4th ckpt
|
|
# it will remove the oldest one to keep the number of total ckpts as 3
|
|
checkpoint_config = dict(interval=10, max_keep_ckpts=3)
|