mirrors
/
fast-reid
mirror of https://github.com/JDAI-CV/fast-reid.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity
fast-reid/projects/FastAttr/configs/Base-attribute.yml

67 lines
1.0 KiB
YAML
Raw Normal View History

update attribute project
2020-09-23 19:45:13 +08:00
MODEL:
update fastreid V1.0
2021-01-18 11:36:38 +08:00
META_ARCHITECTURE: AttrBaseline
update attribute project
2020-09-23 19:45:13 +08:00
BACKBONE:
update fastreid V1.0
2021-01-18 11:36:38 +08:00
NAME: build_resnet_backbone
NORM: BN
DEPTH: 50x
update attribute project
2020-09-23 19:45:13 +08:00
LAST_STRIDE: 2
FEAT_DIM: 2048
WITH_IBN: False
PRETRAIN: True
update fastreid V1.0
2021-01-18 11:36:38 +08:00
PRETRAIN_PATH: /export/home/lxy/.cache/torch/checkpoints/resnet50-19c8e357.pth
update attribute project
2020-09-23 19:45:13 +08:00
HEADS:
update fastreid V1.0
2021-01-18 11:36:38 +08:00
NAME: AttrHead
update attribute project
2020-09-23 19:45:13 +08:00
WITH_BNNECK: True
bugfix for attribute project (#450) Summary: refactor sample weight in attribute recognition; change all options to False in defaults.py and modify yaml files
2021-03-31 17:07:19 +08:00
POOL_LAYER: FastGlobalAvgPool
CLS_LAYER: Linear
update attribute project
2020-09-23 19:45:13 +08:00
NUM_CLASSES: 26
LOSSES:
NAME: ("BinaryCrossEntropyLoss",)
BCE:
WEIGHT_ENABLED: True
SCALE: 1.
INPUT:
update fastreid V1.0
2021-01-18 11:36:38 +08:00
SIZE_TRAIN: [ 256, 192 ]
SIZE_TEST: [ 256, 192 ]
bugfix for attribute project (#450) Summary: refactor sample weight in attribute recognition; change all options to False in defaults.py and modify yaml files
2021-03-31 17:07:19 +08:00
FLIP:
ENABLED: True
PADDING:
ENABLED: True
update attribute project
2020-09-23 19:45:13 +08:00
DATALOADER:
bugfix for attribute project (#450) Summary: refactor sample weight in attribute recognition; change all options to False in defaults.py and modify yaml files
2021-03-31 17:07:19 +08:00
SAMPLER_TRAIN: TrainingSampler
update attribute project
2020-09-23 19:45:13 +08:00
NUM_WORKERS: 8
SOLVER:
update fastreid V1.0
2021-01-18 11:36:38 +08:00
MAX_EPOCH: 30
OPT: SGD
BASE_LR: 0.04
update attribute project
2020-09-23 19:45:13 +08:00
BIAS_LR_FACTOR: 2.
HEADS_LR_FACTOR: 10.
WEIGHT_DECAY: 0.0005
WEIGHT_DECAY_BIAS: 0.0005
update fastreid V1.0
2021-01-18 11:36:38 +08:00
IMS_PER_BATCH: 256
update attribute project
2020-09-23 19:45:13 +08:00
update fastreid V1.0
2021-01-18 11:36:38 +08:00
NESTEROV: False
SCHED: MultiStepLR
STEPS: [ 15, 20, 25 ]
update attribute project
2020-09-23 19:45:13 +08:00
update fastreid V1.0
2021-01-18 11:36:38 +08:00
WARMUP_FACTOR: 0.1
make lr warmup by iter Summary: change warmup way by iter not by epoch, which will make it more flexible when training small epochs
2021-01-22 11:17:21 +08:00
WARMUP_ITERS: 1000
update attribute project
2020-09-23 19:45:13 +08:00
CHECKPOINT_PERIOD: 10
TEST:
EVAL_PERIOD: 10
IMS_PER_BATCH: 256
CUDNN_BENCHMARK: True