NanoDet-Plus⚡Super fast and lightweight anchor-free object detection model. 🔥Only 980 KB(int8) / 1.8MB (fp16) and run 97FPS on cellphone🔥
5.63k
stars
1.03k
forks
source link
Encounter issue while attempting to train Nanodet-plus-m_320 #563
Open
AlphaIkaros2 opened 3 months ago
My configuration file:
nanodet-plus-m_320
COCO mAP(0.5:0.95) = 0.270
AP_50 = 0.418
AP_75 = 0.281
AP_small = 0.083
AP_m = 0.278
AP_l = 0.451
save_dir: workspace/nanodet-plus-m_320 model: weight_averager: name: ExpMovingAverager decay: 0.9998 arch: name: NanoDetPlus detach_epoch: 10 backbone: name: ShuffleNetV2 model_size: 1.0x out_stages: [2,3,4] activation: LeakyReLU fpn: name: GhostPAN in_channels: [116, 232, 464] out_channels: 96 kernel_size: 5 num_extra_level: 1 use_depthwise: True activation: LeakyReLU head: name: NanoDetPlusHead num_classes: 15 input_channel: 96 feat_channels: 96 stacked_convs: 2 kernel_size: 5 strides: [8, 16, 32, 64] activation: LeakyReLU reg_max: 7 norm_cfg: type: BN loss: loss_qfl: name: QualityFocalLoss use_sigmoid: True beta: 2.0 loss_weight: 1.0 loss_dfl: name: DistributionFocalLoss loss_weight: 0.25 loss_bbox: name: GIoULoss loss_weight: 2.0
Auxiliary head, only use in training time.
data: train: name: CocoDataset img_path: /content/drive/MyDrive/Nanodet/BFMC2024_2-3/train ann_path: /content/drive/MyDrive/Nanodet/BFMC2024_2-3/train/_annotations.coco.json input_size: [320,320] #[w,h] keep_ratio: False pipeline: perspective: 0.0 scale: [0.6, 1.4] stretch: [[0.8, 1.2], [0.8, 1.2]] rotation: 0 shear: 0 translate: 0.2 flip: 0.5 brightness: 0.2 contrast: [0.6, 1.4] saturation: [0.5, 1.2] normalize: [[103.53, 116.28, 123.675], [57.375, 57.12, 58.395]] val: name: CocoDataset img_path: /content/drive/MyDrive/Nanodet/BFMC2024_2-3/valid ann_path: /content/drive/MyDrive/Nanodet/BFMC2024_2-3/valid/_annotations.coco.json input_size: [320,320] #[w,h] keep_ratio: False pipeline: normalize: [[103.53, 116.28, 123.675], [57.375, 57.12, 58.395]] device: gpu_ids: [0] # Set like [0, 1, 2, 3] if you have multi-GPUs workers_per_gpu: 10 batchsize_per_gpu: 96 precision: 32 # set to 16 to use AMP training schedule:
resume:
load_model:
optimizer: name: AdamW lr: 0.001 weight_decay: 0.05 warmup: name: linear steps: 500 ratio: 0.0001 total_epochs: 50 lr_schedule: name: CosineAnnealingLR T_max: 300 eta_min: 0.00005 val_intervals: 10 grad_clip: 35 evaluator: name: CocoDetectionEvaluator save_key: mAP log: interval: 50
class_names: ['car','crosswalk','highway_entry' ,'highway_exit' ,'intersection' ,'no_entry' ,'onewayroad' ,'parking' ,'pedestrian' ,'priority' ,'roundabout' ,'stop' ,'trafficlight_green' ,'trafficlight_red' ,'trafficlight_yellow',]
Output: ../aten/src/ATen/native/cuda/ScatterGatherKernel.cu:365: operator(): block: [0,0,0], thread: [0,0,0] Assertion
idx_dim >= 0 && idx_dim < index_size && "index out of bounds"
failed.