gavin/RT-DETR
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

first commit

Browse Source
This commit is contained in:
陈赣
2026-06-03 12:42:47 +08:00
commit ec23799148
339 changed files with 57120 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

38
rtdetrv2_pytorch/configs/rtdetrv2/include/dataloader.yml Normal file
View File

@@ -0,0 +1,38 @@
train_dataloader:
dataset:
transforms:
ops:
- {type: RandomPhotometricDistort, p: 0.5}
- {type: RandomZoomOut, fill: 0}
- {type: RandomIoUCrop, p: 0.8}
- {type: SanitizeBoundingBoxes, min_size: 1}
- {type: RandomHorizontalFlip}
- {type: Resize, size: [640, 640], }
- {type: SanitizeBoundingBoxes, min_size: 1}
- {type: ConvertPILImage, dtype: 'float32', scale: True}
- {type: ConvertBoxes, fmt: 'cxcywh', normalize: True}
policy:
name: stop_epoch
epoch: 71 # epoch in [71, ~) stop `ops`
ops: ['RandomPhotometricDistort', 'RandomZoomOut', 'RandomIoUCrop']
collate_fn:
type: BatchImageCollateFunction
scales: [480, 512, 544, 576, 608, 640, 640, 640, 672, 704, 736, 768, 800]
stop_epoch: 71 # epoch in [71, ~) stop `multiscales`
shuffle: True
total_batch_size: 16 # total batch size equals to 16 (4 * 4)
num_workers: 4
val_dataloader:
dataset:
transforms:
ops:
- {type: Resize, size: [640, 640]}
- {type: ConvertPILImage, dtype: 'float32', scale: True}
shuffle: False
total_batch_size: 32
num_workers: 4

37
rtdetrv2_pytorch/configs/rtdetrv2/include/optimizer.yml Normal file
View File

@@ -0,0 +1,37 @@
use_amp: True
use_ema: True
ema:
type: ModelEMA
decay: 0.9999
warmups: 2000
epoches: 72
clip_max_norm: 0.1
optimizer:
type: AdamW
params:
-
params: '^(?=.*backbone)(?!.*norm).*$'
lr: 0.00001
-
params: '^(?=.*(?:encoder|decoder))(?=.*(?:norm|bn)).*$'
weight_decay: 0.
lr: 0.0001
betas: [0.9, 0.999]
weight_decay: 0.0001
lr_scheduler:
type: MultiStepLR
milestones: [1000]
gamma: 0.1
lr_warmup_scheduler:
type: LinearWarmup
warmup_duration: 2000

83
rtdetrv2_pytorch/configs/rtdetrv2/include/rtdetrv2_r50vd.yml Normal file
View File

@@ -0,0 +1,83 @@
task: detection
model: RTDETR
criterion: RTDETRCriterionv2
postprocessor: RTDETRPostProcessor
use_focal_loss: True
eval_spatial_size: [640, 640] # h w
RTDETR:
backbone: PResNet
encoder: HybridEncoder
decoder: RTDETRTransformerv2
PResNet:
depth: 50
variant: d
freeze_at: 0
return_idx: [1, 2, 3]
num_stages: 4
freeze_norm: True
pretrained: True
HybridEncoder:
in_channels: [512, 1024, 2048]
feat_strides: [8, 16, 32]
# intra
hidden_dim: 256
use_encoder_idx: [2]
num_encoder_layers: 1
nhead: 8
dim_feedforward: 1024
dropout: 0.
enc_act: 'gelu'
# cross
expansion: 1.0
depth_mult: 1
act: 'silu'
RTDETRTransformerv2:
feat_channels: [256, 256, 256]
feat_strides: [8, 16, 32]
hidden_dim: 256
num_levels: 3
num_layers: 6
num_queries: 300
num_denoising: 100
label_noise_ratio: 0.5
box_noise_scale: 1.0 # 1.0 0.4
eval_idx: -1
# NEW
num_points: [4, 4, 4] # [3,3,3] [2,2,2]
cross_attn_method: default # default, discrete
query_select_method: default # default, agnostic
RTDETRPostProcessor:
num_top_queries: 300
RTDETRCriterionv2:
weight_dict: {loss_vfl: 1, loss_bbox: 5, loss_giou: 2,}
losses: ['vfl', 'boxes', ]
alpha: 0.75
gamma: 2.0
matcher:
type: HungarianMatcher
weight_dict: {cost_class: 2, cost_bbox: 5, cost_giou: 2}
alpha: 0.25
gamma: 2.0

50
rtdetrv2_pytorch/configs/rtdetrv2/rtdetrv2_hgnetv2_h_6x_coco.yml Normal file
View File

@@ -0,0 +1,50 @@
__include__: [
'../dataset/coco_detection.yml',
'../runtime.yml',
'./include/dataloader.yml',
'./include/optimizer.yml',
'./include/rtdetrv2_r50vd.yml',
]
output_dir: ./output/rtdetrv2_hgnetv2_h_6x_coco
RTDETR:
backbone: HGNetv2
HGNetv2:
name: 'H'
return_idx: [1, 2, 3]
freeze_at: 0
freeze_norm: True
pretrained: True
HybridEncoder:
# intra
hidden_dim: 512
dim_feedforward: 2048
num_encoder_layers: 2
RTDETRTransformerv2:
feat_channels: [512, 512, 512]
optimizer:
type: AdamW
params:
-
params: '^(?=.*backbone)(?!.*norm|bn).*$'
lr: 0.000005
-
params: '^(?=.*(?:encoder|decoder))(?=.*(?:norm|bn)).*$'
weight_decay: 0.
lr: 0.0001
betas: [0.9, 0.999]
weight_decay: 0.0001

38
rtdetrv2_pytorch/configs/rtdetrv2/rtdetrv2_hgnetv2_l_6x_coco.yml Normal file
View File

@@ -0,0 +1,38 @@
__include__: [
'../dataset/coco_detection.yml',
'../runtime.yml',
'./include/dataloader.yml',
'./include/optimizer.yml',
'./include/rtdetrv2_r50vd.yml',
]
output_dir: ./output/rtdetrv2_hgnetv2_l_6x_coco
RTDETR:
backbone: HGNetv2
HGNetv2:
name: 'L'
return_idx: [1, 2, 3]
freeze_at: 0
freeze_norm: True
pretrained: True
optimizer:
type: AdamW
params:
-
params: '^(?=.*backbone)(?!.*norm|bn).*$'
lr: 0.000005
-
params: '^(?=.*(?:encoder|decoder))(?=.*(?:norm|bn)).*$'
weight_decay: 0.
lr: 0.0001
betas: [0.9, 0.999]
weight_decay: 0.0001

50
rtdetrv2_pytorch/configs/rtdetrv2/rtdetrv2_hgnetv2_x_6x_coco.yml Normal file
View File

@@ -0,0 +1,50 @@
__include__: [
'../dataset/coco_detection.yml',
'../runtime.yml',
'./include/dataloader.yml',
'./include/optimizer.yml',
'./include/rtdetrv2_r50vd.yml',
]
output_dir: ./output/rtdetrv2_hgnetv2_x_6x_coco
RTDETR:
backbone: HGNetv2
HGNetv2:
name: 'X'
return_idx: [1, 2, 3]
freeze_at: 0
freeze_norm: True
pretrained: True
HybridEncoder:
# intra
hidden_dim: 384
dim_feedforward: 2048
RTDETRTransformerv2:
feat_channels: [384, 384, 384]
optimizer:
type: AdamW
params:
-
params: '^(?=.*backbone)(?!.*norm|bn).*$'
lr: 0.000001
-
params: '^(?=.*(?:encoder|decoder))(?=.*(?:norm|bn)).*$'
weight_decay: 0.
lr: 0.0001
betas: [0.9, 0.999]
weight_decay: 0.0001

40
rtdetrv2_pytorch/configs/rtdetrv2/rtdetrv2_r101vd_6x_coco.yml Normal file
View File

@@ -0,0 +1,40 @@
__include__: [
'../dataset/coco_detection.yml',
'../runtime.yml',
'./include/dataloader.yml',
'./include/optimizer.yml',
'./include/rtdetrv2_r50vd.yml',
]
output_dir: ./output/rtdetrv2_r101vd_6x_coco
PResNet:
depth: 101
HybridEncoder:
# intra
hidden_dim: 384
dim_feedforward: 2048
RTDETRTransformerv2:
feat_channels: [384, 384, 384]
optimizer:
type: AdamW
params:
-
params: '^(?=.*backbone)(?!.*norm|bn).*$'
lr: 0.000001
-
params: '^(?=.*(?:encoder|decoder))(?=.*(?:norm|bn)).*$'
weight_decay: 0.
lr: 0.0001
betas: [0.9, 0.999]
weight_decay: 0.0001

46
rtdetrv2_pytorch/configs/rtdetrv2/rtdetrv2_r18vd_120e_coco.yml Normal file
View File

@@ -0,0 +1,46 @@
__include__: [
'../dataset/coco_detection.yml',
'../runtime.yml',
'./include/dataloader.yml',
'./include/optimizer.yml',
'./include/rtdetrv2_r50vd.yml',
]
output_dir: ./output/rtdetrv2_r18vd_120e_coco
PResNet:
depth: 18
freeze_at: -1
freeze_norm: False
pretrained: True
HybridEncoder:
in_channels: [128, 256, 512]
hidden_dim: 256
expansion: 0.5
RTDETRTransformerv2:
num_layers: 3
epoches: 120
optimizer:
type: AdamW
params:
-
params: '^(?=.*(?:norm|bn)).*$'
weight_decay: 0.
train_dataloader:
dataset:
transforms:
policy:
epoch: 117
collate_fn:
scales: ~

46
rtdetrv2_pytorch/configs/rtdetrv2/rtdetrv2_r18vd_120e_voc.yml Normal file
View File

@@ -0,0 +1,46 @@
__include__: [
'../dataset/voc_detection.yml',
'../runtime.yml',
'./include/dataloader.yml',
'./include/optimizer.yml',
'./include/rtdetrv2_r50vd.yml',
]
output_dir: ./output/rtdetrv2_r18vd_120e_voc
PResNet:
depth: 18
freeze_at: -1
freeze_norm: False
pretrained: True
HybridEncoder:
in_channels: [128, 256, 512]
hidden_dim: 256
expansion: 0.5
RTDETRTransformerv2:
num_layers: 3
epoches: 120
optimizer:
type: AdamW
params:
-
params: '^(?=.*(?:norm|bn)).*$'
weight_decay: 0.
train_dataloader:
dataset:
transforms:
policy:
epoch: 117
collate_fn:
scales: ~
total_batch_size: 32

49
rtdetrv2_pytorch/configs/rtdetrv2/rtdetrv2_r18vd_dsp_3x_coco.yml Normal file
View File

@@ -0,0 +1,49 @@
__include__: [
'../dataset/coco_detection.yml',
'../runtime.yml',
'./include/dataloader.yml',
'./include/optimizer.yml',
'./include/rtdetrv2_r50vd.yml',
]
tuning: https://github.com/lyuwenyu/storage/releases/download/v0.1/rtdetrv2_r18vd_120e_coco.pth
output_dir: ./output/rtdetrv2_r18vd_dsp_3x_coco
PResNet:
depth: 18
freeze_at: -1
freeze_norm: False
pretrained: True
HybridEncoder:
in_channels: [128, 256, 512]
hidden_dim: 256
expansion: 0.5
RTDETRTransformerv2:
num_layers: 3
num_points: [4, 4, 4]
cross_attn_method: discrete
epoches: 36
optimizer:
type: AdamW
params:
-
params: '^(?=.*(?:norm|bn)).*$'
weight_decay: 0.
train_dataloader:
dataset:
transforms:
policy:
epoch: 33
collate_fn:
scales: ~

47
rtdetrv2_pytorch/configs/rtdetrv2/rtdetrv2_r18vd_sp1_120e_coco.yml Normal file
View File

@@ -0,0 +1,47 @@
__include__: [
'../dataset/coco_detection.yml',
'../runtime.yml',
'./include/dataloader.yml',
'./include/optimizer.yml',
'./include/rtdetrv2_r50vd.yml',
]
output_dir: ./output/rtdetrv2_r18vd_sp1_120e_coco
PResNet:
depth: 18
freeze_at: -1
freeze_norm: False
pretrained: True
HybridEncoder:
in_channels: [128, 256, 512]
hidden_dim: 256
expansion: 0.5
RTDETRTransformerv2:
num_layers: 3
num_points: [1, 1, 1]
epoches: 120
optimizer:
type: AdamW
params:
-
params: '^(?=.*(?:norm|bn)).*$'
weight_decay: 0.
train_dataloader:
dataset:
transforms:
policy:
epoch: 117
collate_fn:
scales: ~

47
rtdetrv2_pytorch/configs/rtdetrv2/rtdetrv2_r18vd_sp2_120e_coco.yml Normal file
View File

@@ -0,0 +1,47 @@
__include__: [
'../dataset/coco_detection.yml',
'../runtime.yml',
'./include/dataloader.yml',
'./include/optimizer.yml',
'./include/rtdetrv2_r50vd.yml',
]
output_dir: ./output/rtdetrv2_r18vd_sp2_120e_coco
PResNet:
depth: 18
freeze_at: -1
freeze_norm: False
pretrained: True
HybridEncoder:
in_channels: [128, 256, 512]
hidden_dim: 256
expansion: 0.5
RTDETRTransformerv2:
num_layers: 3
num_points: [2, 2, 2]
epoches: 120
optimizer:
type: AdamW
params:
-
params: '^(?=.*(?:norm|bn)).*$'
weight_decay: 0.
train_dataloader:
dataset:
transforms:
policy:
epoch: 117
collate_fn:
scales: ~

47
rtdetrv2_pytorch/configs/rtdetrv2/rtdetrv2_r18vd_sp3_120e_coco.yml Normal file
View File

@@ -0,0 +1,47 @@
__include__: [
'../dataset/coco_detection.yml',
'../runtime.yml',
'./include/dataloader.yml',
'./include/optimizer.yml',
'./include/rtdetrv2_r50vd.yml',
]
output_dir: ./output/rtdetrv2_r18vd_sp3_120e_coco
PResNet:
depth: 18
freeze_at: -1
freeze_norm: False
pretrained: True
HybridEncoder:
in_channels: [128, 256, 512]
hidden_dim: 256
expansion: 0.5
RTDETRTransformerv2:
num_layers: 3
num_points: [3, 3, 3]
epoches: 120
optimizer:
type: AdamW
params:
-
params: '^(?=.*(?:norm|bn)).*$'
weight_decay: 0.
train_dataloader:
dataset:
transforms:
policy:
epoch: 117
collate_fn:
scales: ~

57
rtdetrv2_pytorch/configs/rtdetrv2/rtdetrv2_r34vd_120e_coco.yml Normal file
View File

@@ -0,0 +1,57 @@
__include__: [
'../dataset/coco_detection.yml',
'../runtime.yml',
'./include/dataloader.yml',
'./include/optimizer.yml',
'./include/rtdetrv2_r50vd.yml',
]
output_dir: ./output/rtdetrv2_r34vd_120e_coco
PResNet:
depth: 34
freeze_at: -1
freeze_norm: False
pretrained: True
HybridEncoder:
in_channels: [128, 256, 512]
hidden_dim: 256
expansion: 0.5
RTDETRTransformerv2:
num_layers: 4
epoches: 120
optimizer:
type: AdamW
params:
-
params: '^(?=.*backbone)(?!.*norm|bn).*$'
lr: 0.00005
-
params: '^(?=.*backbone)(?=.*norm|bn).*$'
lr: 0.00005
weight_decay: 0.
-
params: '^(?=.*(?:encoder|decoder))(?=.*(?:norm|bn|bias)).*$'
weight_decay: 0.
lr: 0.0001
betas: [0.9, 0.999]
weight_decay: 0.0001
train_dataloader:
dataset:
transforms:
policy:
epoch: 117
collate_fn:
stop_epoch: 117

59
rtdetrv2_pytorch/configs/rtdetrv2/rtdetrv2_r34vd_dsp_1x_coco.yml Normal file
View File

@@ -0,0 +1,59 @@
__include__: [
'../dataset/coco_detection.yml',
'../runtime.yml',
'./include/dataloader.yml',
'./include/optimizer.yml',
'./include/rtdetrv2_r50vd.yml',
]
tuning: https://github.com/lyuwenyu/storage/releases/download/v0.1/rtdetrv2_r34vd_120e_coco_ema.pth
output_dir: ./output/rtdetrv2_r34vd_dsp_1x_coco
PResNet:
depth: 34
freeze_at: -1
freeze_norm: False
pretrained: True
HybridEncoder:
in_channels: [128, 256, 512]
hidden_dim: 256
expansion: 0.5
RTDETRTransformerv2:
num_layers: 4
cross_attn_method: discrete
epoches: 12
optimizer:
type: AdamW
params:
-
params: '^(?=.*backbone)(?!.*norm|bn).*$'
lr: 0.00005
-
params: '^(?=.*backbone)(?=.*norm|bn).*$'
lr: 0.00005
weight_decay: 0.
-
params: '^(?=.*(?:encoder|decoder))(?=.*(?:norm|bn|bias)).*$'
weight_decay: 0.
lr: 0.0001
betas: [0.9, 0.999]
weight_decay: 0.0001
train_dataloader:
dataset:
transforms:
policy:
epoch: 10
collate_fn:
stop_epoch: 10

27
rtdetrv2_pytorch/configs/rtdetrv2/rtdetrv2_r50vd_6x_coco.yml Normal file
View File

@@ -0,0 +1,27 @@
__include__: [
'../dataset/coco_detection.yml',
'../runtime.yml',
'./include/dataloader.yml',
'./include/optimizer.yml',
'./include/rtdetrv2_r50vd.yml',
]
output_dir: ./output/rtdetrv2_r50vd_6x_coco
optimizer:
type: AdamW
params:
-
params: '^(?=.*backbone)(?!.*norm).*$'
lr: 0.00001
-
params: '^(?=.*(?:encoder|decoder))(?=.*(?:norm|bn)).*$'
weight_decay: 0.
lr: 0.0001
betas: [0.9, 0.999]
weight_decay: 0.0001

27
rtdetrv2_pytorch/configs/rtdetrv2/rtdetrv2_r50vd_dsp_1x_coco.yml Normal file
View File

@@ -0,0 +1,27 @@
__include__: [
'../dataset/coco_detection.yml',
'../runtime.yml',
'./include/dataloader.yml',
'./include/optimizer.yml',
'./include/rtdetrv2_r50vd.yml',
]
tuning: https://github.com/lyuwenyu/storage/releases/download/v0.1/rtdetrv2_r50vd_6x_coco_ema.pth
output_dir: ./output/rtdetrv2_r50vd_dsp_1x_coco
RTDETRTransformerv2:
cross_attn_method: discrete
epoches: 12
train_dataloader:
dataset:
transforms:
policy:
epoch: 10
collate_fn:
stop_epoch: 10

43
rtdetrv2_pytorch/configs/rtdetrv2/rtdetrv2_r50vd_m_7x_coco.yml Normal file
View File

@@ -0,0 +1,43 @@
__include__: [
'../dataset/coco_detection.yml',
'../runtime.yml',
'./include/dataloader.yml',
'./include/optimizer.yml',
'./include/rtdetrv2_r50vd.yml',
]
output_dir: ./output/rtdetrv2_r50vd_m_6x_coco
HybridEncoder:
expansion: 0.5
RTDETRTransformerv2:
eval_idx: 2 # use 3th decoder layer to eval
epoches: 84
optimizer:
type: AdamW
params:
-
params: '^(?=.*backbone)(?!.*norm).*$'
lr: 0.00001
-
params: '^(?=.*(?:encoder|decoder))(?=.*(?:norm|bn)).*$'
weight_decay: 0.
lr: 0.0001
betas: [0.9, 0.999]
weight_decay: 0.0001
train_dataloader:
dataset:
transforms:
policy:
epoch: 81
collate_fn:
stop_epoch: 81

44
rtdetrv2_pytorch/configs/rtdetrv2/rtdetrv2_r50vd_m_dsp_3x_coco.yml Normal file
View File

@@ -0,0 +1,44 @@
__include__: [
'../dataset/coco_detection.yml',
'../runtime.yml',
'./include/dataloader.yml',
'./include/optimizer.yml',
'./include/rtdetrv2_r50vd.yml',
]
output_dir: ./output/rtdetrv2_r50vd_m_dsp_3x_coco
tuning: https://github.com/lyuwenyu/storage/releases/download/v0.1/rtdetrv2_r50vd_m_7x_coco_ema.pth
HybridEncoder:
expansion: 0.5
RTDETRTransformerv2:
eval_idx: 2 # use 3th decoder layer to eval
cross_attn_method: discrete
epoches: 36
optimizer:
type: AdamW
params:
-
params: '^(?=.*backbone)(?!.*norm).*$'
lr: 0.00001
-
params: '^(?=.*(?:encoder|decoder))(?=.*(?:norm|bn)).*$'
weight_decay: 0.
lr: 0.0001
betas: [0.9, 0.999]
weight_decay: 0.0001
train_dataloader:
dataset:
transforms:
policy:
epoch: 33
collate_fn:
stop_epoch: 33