[ 2025-06-08 17:13 ] Loading ResNet50 weights and modifying its structure.
[ 2025-06-08 17:13 ] Model load finished: model.ISTANet.Model
[ 2025-06-08 17:15 ] Loading ResNet50 weights and modifying its structure.
[ 2025-06-08 17:15 ] Model load finished: model.ISTANet.Model
[ 2025-06-08 17:15 ] Data load finished
[ 2025-06-08 17:15 ] Optimizer load finished: AdamW
[ 2025-06-08 17:15 ] Loading ResNet50 weights and modifying its structure.
[ 2025-06-08 17:15 ] Model load finished: model.ISTANet.Model
[ 2025-06-08 17:15 ] Data load finished
[ 2025-06-08 17:15 ] Optimizer load finished: AdamW
[ 2025-06-08 17:15 ] action_loss_weight: 1
[ 2025-06-08 17:15 ] base_lr: 0.001
[ 2025-06-08 17:15 ] batch_size: 16
[ 2025-06-08 17:15 ] config: ./config/fpha/fpha.yaml
[ 2025-06-08 17:15 ] cuda_visible_device: 0
[ 2025-06-08 17:15 ] device: [0]
[ 2025-06-08 17:15 ] eval_interval: 2
[ 2025-06-08 17:15 ] feat_loss_weight: 1.2
[ 2025-06-08 17:15 ] feeder: feeders.feeder_fpha.Feeder
[ 2025-06-08 17:15 ] ignore_weights: []
[ 2025-06-08 17:15 ] loss: LabelSmoothingCrossEntropy
[ 2025-06-08 17:15 ] loss_args: {'smoothing': 0.15, 'temperature': 1.0}
[ 2025-06-08 17:15 ] lr_decay_rate: 0.5
[ 2025-06-08 17:15 ] model: model.ISTANet.Model
[ 2025-06-08 17:15 ] model_args: {'window_size': [20, 1, 3], 'num_frames': 120, 'num_joints': 21, 'num_persons': 2, 'num_channels': 3, 'num_classes': 45, 'num_objs': 26, 'num_verbs': 27, 'num_heads': 4, 'kernel_size': [3, 5], 'use_pes': True, 'config': [[64, 64, 16], [64, 64, 16], [64, 128, 32], [128, 128, 32], [128, 256, 64], [256, 256, 64], [256, 256, 64], [256, 256, 64]]}
[ 2025-06-08 17:15 ] nesterov: True
[ 2025-06-08 17:15 ] noun_loss_weight: 1.5
[ 2025-06-08 17:15 ] num_epoch: 150
[ 2025-06-08 17:15 ] num_worker: 8
[ 2025-06-08 17:15 ] optimizer: AdamW
[ 2025-06-08 17:15 ] optimizer_betas: [0.9, 0.999]
[ 2025-06-08 17:15 ] print_log: True
[ 2025-06-08 17:15 ] run_mode: train
[ 2025-06-08 17:15 ] save_epoch: 80
[ 2025-06-08 17:15 ] save_score: False
[ 2025-06-08 17:15 ] seed: 1
[ 2025-06-08 17:15 ] show_topk: [1, 5]
[ 2025-06-08 17:15 ] start_epoch: 0
[ 2025-06-08 17:15 ] step: [50, 80, 110, 130]
[ 2025-06-08 17:15 ] test_batch_size: 16
[ 2025-06-08 17:15 ] test_feeder_args: {'data_path': 'data/fpha/fpha_pth', 'split': 'val', 'debug': False, 'window_size': 120, 'p_interval': [0.95], 'vel': False, 'bone': False, 'use_rgb': True}
[ 2025-06-08 17:15 ] train_feeder_args: {'data_path': 'data/fpha/fpha_pth', 'split': 'train', 'debug': False, 'random_choose': False, 'random_shift': False, 'random_move': False, 'window_size': 120, 'normalization': False, 'random_rot': True, 'p_interval': [0.5, 1], 'vel': False, 'bone': False, 'entity_rearrangement': False, 'use_rgb': True}
[ 2025-06-08 17:15 ] verb_loss_weight: 1
[ 2025-06-08 17:15 ] warm_up_epoch: 5
[ 2025-06-08 17:15 ] weight_decay: 0.0015
[ 2025-06-08 17:15 ] weights: None
[ 2025-06-08 17:15 ] work_dir: ./exp/fpha
[ 2025-06-08 17:15 ] # Parameters: 33605890
[ 2025-06-08 17:15 ] ###***************start training***************###
[ 2025-06-08 17:15 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 17:18 ] Loading ResNet50 weights and modifying its structure.
[ 2025-06-08 17:18 ] Model load finished: model.ISTANet.Model
[ 2025-06-08 17:18 ] Data load finished
[ 2025-06-08 17:18 ] Optimizer load finished: AdamW
[ 2025-06-08 17:18 ] action_loss_weight: 1
[ 2025-06-08 17:18 ] base_lr: 0.001
[ 2025-06-08 17:18 ] batch_size: 16
[ 2025-06-08 17:18 ] config: ./config/fpha/fpha.yaml
[ 2025-06-08 17:18 ] cuda_visible_device: 0
[ 2025-06-08 17:18 ] device: [0]
[ 2025-06-08 17:18 ] eval_interval: 2
[ 2025-06-08 17:18 ] feat_loss_weight: 1.2
[ 2025-06-08 17:18 ] feeder: feeders.feeder_fpha.Feeder
[ 2025-06-08 17:18 ] ignore_weights: []
[ 2025-06-08 17:18 ] loss: LabelSmoothingCrossEntropy
[ 2025-06-08 17:18 ] loss_args: {'smoothing': 0.15, 'temperature': 1.0}
[ 2025-06-08 17:18 ] lr_decay_rate: 0.5
[ 2025-06-08 17:18 ] model: model.ISTANet.Model
[ 2025-06-08 17:18 ] model_args: {'window_size': [20, 1, 2], 'num_frames': 120, 'num_joints': 21, 'num_persons': 2, 'num_channels': 3, 'num_classes': 45, 'num_objs': 26, 'num_verbs': 27, 'num_heads': 4, 'kernel_size': [3, 5], 'use_pes': True, 'config': [[64, 64, 16], [64, 64, 16], [64, 128, 32], [128, 128, 32], [128, 256, 64], [256, 256, 64], [256, 256, 64], [256, 256, 64]]}
[ 2025-06-08 17:18 ] nesterov: True
[ 2025-06-08 17:18 ] noun_loss_weight: 1.5
[ 2025-06-08 17:18 ] num_epoch: 150
[ 2025-06-08 17:18 ] num_worker: 8
[ 2025-06-08 17:18 ] optimizer: AdamW
[ 2025-06-08 17:18 ] optimizer_betas: [0.9, 0.999]
[ 2025-06-08 17:18 ] print_log: True
[ 2025-06-08 17:18 ] run_mode: train
[ 2025-06-08 17:18 ] save_epoch: 80
[ 2025-06-08 17:18 ] save_score: False
[ 2025-06-08 17:18 ] seed: 1
[ 2025-06-08 17:18 ] show_topk: [1, 5]
[ 2025-06-08 17:18 ] start_epoch: 0
[ 2025-06-08 17:18 ] step: [50, 80, 110, 130]
[ 2025-06-08 17:18 ] test_batch_size: 16
[ 2025-06-08 17:18 ] test_feeder_args: {'data_path': 'data/fpha/fpha_pth', 'split': 'val', 'debug': False, 'window_size': 120, 'p_interval': [0.95], 'vel': False, 'bone': False, 'use_rgb': True}
[ 2025-06-08 17:18 ] train_feeder_args: {'data_path': 'data/fpha/fpha_pth', 'split': 'train', 'debug': False, 'random_choose': False, 'random_shift': False, 'random_move': False, 'window_size': 120, 'normalization': False, 'random_rot': True, 'p_interval': [0.5, 1], 'vel': False, 'bone': False, 'entity_rearrangement': False, 'use_rgb': True}
[ 2025-06-08 17:18 ] verb_loss_weight: 1
[ 2025-06-08 17:18 ] warm_up_epoch: 5
[ 2025-06-08 17:18 ] weight_decay: 0.0015
[ 2025-06-08 17:18 ] weights: None
[ 2025-06-08 17:18 ] work_dir: ./exp/fpha
[ 2025-06-08 17:18 ] # Parameters: 33605890
[ 2025-06-08 17:18 ] ###***************start training***************###
[ 2025-06-08 17:18 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 17:20 ] Loading ResNet50 weights and modifying its structure.
[ 2025-06-08 17:20 ] Model load finished: model.ISTANet.Model
[ 2025-06-08 17:20 ] Data load finished
[ 2025-06-08 17:20 ] Optimizer load finished: AdamW
[ 2025-06-08 17:20 ] action_loss_weight: 1
[ 2025-06-08 17:20 ] base_lr: 0.001
[ 2025-06-08 17:20 ] batch_size: 8
[ 2025-06-08 17:20 ] config: ./config/fpha/fpha.yaml
[ 2025-06-08 17:20 ] cuda_visible_device: 0
[ 2025-06-08 17:20 ] device: [0]
[ 2025-06-08 17:20 ] eval_interval: 2
[ 2025-06-08 17:20 ] feat_loss_weight: 1.2
[ 2025-06-08 17:20 ] feeder: feeders.feeder_fpha.Feeder
[ 2025-06-08 17:20 ] ignore_weights: []
[ 2025-06-08 17:20 ] loss: LabelSmoothingCrossEntropy
[ 2025-06-08 17:20 ] loss_args: {'smoothing': 0.15, 'temperature': 1.0}
[ 2025-06-08 17:20 ] lr_decay_rate: 0.5
[ 2025-06-08 17:20 ] model: model.ISTANet.Model
[ 2025-06-08 17:20 ] model_args: {'window_size': [20, 1, 2], 'num_frames': 120, 'num_joints': 21, 'num_persons': 2, 'num_channels': 3, 'num_classes': 45, 'num_objs': 26, 'num_verbs': 27, 'num_heads': 4, 'kernel_size': [3, 5], 'use_pes': True, 'config': [[64, 64, 16], [64, 64, 16], [64, 128, 32], [128, 128, 32], [128, 256, 64], [256, 256, 64], [256, 256, 64], [256, 256, 64]]}
[ 2025-06-08 17:20 ] nesterov: True
[ 2025-06-08 17:20 ] noun_loss_weight: 1.5
[ 2025-06-08 17:20 ] num_epoch: 150
[ 2025-06-08 17:20 ] num_worker: 8
[ 2025-06-08 17:20 ] optimizer: AdamW
[ 2025-06-08 17:20 ] optimizer_betas: [0.9, 0.999]
[ 2025-06-08 17:20 ] print_log: True
[ 2025-06-08 17:20 ] run_mode: train
[ 2025-06-08 17:20 ] save_epoch: 80
[ 2025-06-08 17:20 ] save_score: False
[ 2025-06-08 17:20 ] seed: 1
[ 2025-06-08 17:20 ] show_topk: [1, 5]
[ 2025-06-08 17:20 ] start_epoch: 0
[ 2025-06-08 17:20 ] step: [50, 80, 110, 130]
[ 2025-06-08 17:20 ] test_batch_size: 16
[ 2025-06-08 17:20 ] test_feeder_args: {'data_path': 'data/fpha/fpha_pth', 'split': 'val', 'debug': False, 'window_size': 120, 'p_interval': [0.95], 'vel': False, 'bone': False, 'use_rgb': True}
[ 2025-06-08 17:20 ] train_feeder_args: {'data_path': 'data/fpha/fpha_pth', 'split': 'train', 'debug': False, 'random_choose': False, 'random_shift': False, 'random_move': False, 'window_size': 120, 'normalization': False, 'random_rot': True, 'p_interval': [0.5, 1], 'vel': False, 'bone': False, 'entity_rearrangement': False, 'use_rgb': True}
[ 2025-06-08 17:20 ] verb_loss_weight: 1
[ 2025-06-08 17:20 ] warm_up_epoch: 5
[ 2025-06-08 17:20 ] weight_decay: 0.0015
[ 2025-06-08 17:20 ] weights: None
[ 2025-06-08 17:20 ] work_dir: ./exp/fpha
[ 2025-06-08 17:20 ] # Parameters: 33605890
[ 2025-06-08 17:20 ] ###***************start training***************###
[ 2025-06-08 17:20 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 17:22 ] Loading ResNet50 weights and modifying its structure.
[ 2025-06-08 17:22 ] Model load finished: model.ISTANet.Model
[ 2025-06-08 17:22 ] Data load finished
[ 2025-06-08 17:22 ] Optimizer load finished: AdamW
[ 2025-06-08 17:22 ] action_loss_weight: 1
[ 2025-06-08 17:22 ] base_lr: 0.001
[ 2025-06-08 17:22 ] batch_size: 4
[ 2025-06-08 17:22 ] config: ./config/fpha/fpha.yaml
[ 2025-06-08 17:22 ] cuda_visible_device: 0
[ 2025-06-08 17:22 ] device: [0]
[ 2025-06-08 17:22 ] eval_interval: 2
[ 2025-06-08 17:22 ] feat_loss_weight: 1.2
[ 2025-06-08 17:22 ] feeder: feeders.feeder_fpha.Feeder
[ 2025-06-08 17:22 ] ignore_weights: []
[ 2025-06-08 17:22 ] loss: LabelSmoothingCrossEntropy
[ 2025-06-08 17:22 ] loss_args: {'smoothing': 0.15, 'temperature': 1.0}
[ 2025-06-08 17:22 ] lr_decay_rate: 0.5
[ 2025-06-08 17:22 ] model: model.ISTANet.Model
[ 2025-06-08 17:22 ] model_args: {'window_size': [20, 1, 2], 'num_frames': 120, 'num_joints': 21, 'num_persons': 2, 'num_channels': 3, 'num_classes': 45, 'num_objs': 26, 'num_verbs': 27, 'num_heads': 4, 'kernel_size': [3, 5], 'use_pes': True, 'config': [[64, 64, 16], [64, 64, 16], [64, 128, 32], [128, 128, 32], [128, 256, 64], [256, 256, 64], [256, 256, 64], [256, 256, 64]]}
[ 2025-06-08 17:22 ] nesterov: True
[ 2025-06-08 17:22 ] noun_loss_weight: 1.5
[ 2025-06-08 17:22 ] num_epoch: 150
[ 2025-06-08 17:22 ] num_worker: 8
[ 2025-06-08 17:22 ] optimizer: AdamW
[ 2025-06-08 17:22 ] optimizer_betas: [0.9, 0.999]
[ 2025-06-08 17:22 ] print_log: True
[ 2025-06-08 17:22 ] run_mode: train
[ 2025-06-08 17:22 ] save_epoch: 80
[ 2025-06-08 17:22 ] save_score: False
[ 2025-06-08 17:22 ] seed: 1
[ 2025-06-08 17:22 ] show_topk: [1, 5]
[ 2025-06-08 17:22 ] start_epoch: 0
[ 2025-06-08 17:22 ] step: [50, 80, 110, 130]
[ 2025-06-08 17:22 ] test_batch_size: 8
[ 2025-06-08 17:22 ] test_feeder_args: {'data_path': 'data/fpha/fpha_pth', 'split': 'val', 'debug': False, 'window_size': 120, 'p_interval': [0.95], 'vel': False, 'bone': False, 'use_rgb': True}
[ 2025-06-08 17:22 ] train_feeder_args: {'data_path': 'data/fpha/fpha_pth', 'split': 'train', 'debug': False, 'random_choose': False, 'random_shift': False, 'random_move': False, 'window_size': 120, 'normalization': False, 'random_rot': True, 'p_interval': [0.5, 1], 'vel': False, 'bone': False, 'entity_rearrangement': False, 'use_rgb': True}
[ 2025-06-08 17:22 ] verb_loss_weight: 1
[ 2025-06-08 17:22 ] warm_up_epoch: 5
[ 2025-06-08 17:22 ] weight_decay: 0.0015
[ 2025-06-08 17:22 ] weights: None
[ 2025-06-08 17:22 ] work_dir: ./exp/fpha
[ 2025-06-08 17:22 ] # Parameters: 33605890
[ 2025-06-08 17:22 ] ###***************start training***************###
[ 2025-06-08 17:22 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 17:24 ] training: epoch: 1, loss: 11.3878, top1: 4.33%, lr: 0.000200, obj_loss: 2.9715, verb_loss:3.0901,l1_loss:0.0534, obj_acc:13.3333%, verb_acc:13.5000%
[ 2025-06-08 17:24 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 17:26 ] training: epoch: 2, loss: 10.0461, top1: 6.83%, lr: 0.000400, obj_loss: 2.4498, verb_loss:2.9324,l1_loss:0.0176, obj_acc:23.6667%, verb_acc:14.1667%
[ 2025-06-08 17:29 ] Loading ResNet50 weights and modifying its structure.
[ 2025-06-08 17:29 ] Model load finished: model.ISTANet.Model
[ 2025-06-08 17:29 ] Data load finished
[ 2025-06-08 17:29 ] Optimizer load finished: AdamW
[ 2025-06-08 17:29 ] action_loss_weight: 1
[ 2025-06-08 17:29 ] base_lr: 0.001
[ 2025-06-08 17:29 ] batch_size: 4
[ 2025-06-08 17:29 ] config: ./config/fpha/fpha.yaml
[ 2025-06-08 17:29 ] cuda_visible_device: 0
[ 2025-06-08 17:29 ] device: [0]
[ 2025-06-08 17:29 ] eval_interval: 2
[ 2025-06-08 17:29 ] feat_loss_weight: 1.2
[ 2025-06-08 17:29 ] feeder: feeders.feeder_fpha.Feeder
[ 2025-06-08 17:29 ] ignore_weights: []
[ 2025-06-08 17:29 ] loss: LabelSmoothingCrossEntropy
[ 2025-06-08 17:29 ] loss_args: {'smoothing': 0.15, 'temperature': 1.0}
[ 2025-06-08 17:29 ] lr_decay_rate: 0.5
[ 2025-06-08 17:29 ] model: model.ISTANet.Model
[ 2025-06-08 17:29 ] model_args: {'window_size': [20, 1, 2], 'num_frames': 120, 'num_joints': 21, 'num_persons': 2, 'num_channels': 3, 'num_classes': 45, 'num_objs': 26, 'num_verbs': 27, 'num_heads': 4, 'kernel_size': [3, 5], 'use_pes': True, 'config': [[64, 64, 16], [64, 64, 16], [64, 128, 32], [128, 128, 32], [128, 256, 64], [256, 256, 64], [256, 256, 64], [256, 256, 64]]}
[ 2025-06-08 17:29 ] nesterov: True
[ 2025-06-08 17:29 ] noun_loss_weight: 1.5
[ 2025-06-08 17:29 ] num_epoch: 150
[ 2025-06-08 17:29 ] num_worker: 8
[ 2025-06-08 17:29 ] optimizer: AdamW
[ 2025-06-08 17:29 ] optimizer_betas: [0.9, 0.999]
[ 2025-06-08 17:29 ] print_log: True
[ 2025-06-08 17:29 ] run_mode: train
[ 2025-06-08 17:29 ] save_epoch: 80
[ 2025-06-08 17:29 ] save_score: False
[ 2025-06-08 17:29 ] seed: 1
[ 2025-06-08 17:29 ] show_topk: [1, 5]
[ 2025-06-08 17:29 ] start_epoch: 0
[ 2025-06-08 17:29 ] step: [50, 80, 110, 130]
[ 2025-06-08 17:29 ] test_batch_size: 4
[ 2025-06-08 17:29 ] test_feeder_args: {'data_path': 'data/fpha/fpha_pth', 'split': 'val', 'debug': False, 'window_size': 120, 'p_interval': [0.95], 'vel': False, 'bone': False, 'use_rgb': True}
[ 2025-06-08 17:29 ] train_feeder_args: {'data_path': 'data/fpha/fpha_pth', 'split': 'train', 'debug': False, 'random_choose': False, 'random_shift': False, 'random_move': False, 'window_size': 120, 'normalization': False, 'random_rot': True, 'p_interval': [0.5, 1], 'vel': False, 'bone': False, 'entity_rearrangement': False, 'use_rgb': True}
[ 2025-06-08 17:29 ] verb_loss_weight: 1
[ 2025-06-08 17:29 ] warm_up_epoch: 5
[ 2025-06-08 17:29 ] weight_decay: 0.0015
[ 2025-06-08 17:29 ] weights: None
[ 2025-06-08 17:29 ] work_dir: ./exp/fpha
[ 2025-06-08 17:29 ] # Parameters: 33605890
[ 2025-06-08 17:29 ] ###***************start training***************###
[ 2025-06-08 17:29 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 17:31 ] training: epoch: 1, loss: 11.3878, top1: 4.33%, lr: 0.000200, obj_loss: 2.9715, verb_loss:3.0901,l1_loss:0.0534, obj_acc:13.3333%, verb_acc:13.5000%
[ 2025-06-08 17:31 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 17:33 ] training: epoch: 2, loss: 10.0461, top1: 6.83%, lr: 0.000400, obj_loss: 2.4498, verb_loss:2.9324,l1_loss:0.0176, obj_acc:23.6667%, verb_acc:14.1667%
[ 2025-06-08 17:35 ] evaluating: loss: 10.5443, top1: 9.39%, best_acc: 9.39%,obj_loss: 2.6835,verb_loss:3.0039,l1_loss:0.0094,obj_acc:18.2609%, verb_acc:16.1739%
[ 2025-06-08 17:35 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 17:38 ] training: epoch: 3, loss: 9.6123, top1: 11.50%, lr: 0.000600, obj_loss: 2.3104, verb_loss:2.8505,l1_loss:0.0156, obj_acc:23.5000%, verb_acc:17.1667%
[ 2025-06-08 17:38 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 17:40 ] training: epoch: 4, loss: 9.4647, top1: 13.00%, lr: 0.000800, obj_loss: 2.2471, verb_loss:2.8081,l1_loss:0.0130, obj_acc:25.8333%, verb_acc:21.5000%
[ 2025-06-08 17:42 ] evaluating: loss: 11.2327, top1: 11.30%, best_acc: 11.30%,obj_loss: 3.0762,verb_loss:3.0099,l1_loss:0.0087,obj_acc:23.6522%, verb_acc:16.1739%
[ 2025-06-08 17:42 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 17:44 ] training: epoch: 5, loss: 10.6127, top1: 9.67%, lr: 0.001000, obj_loss: 2.7014, verb_loss:2.9960,l1_loss:0.0093, obj_acc:20.0000%, verb_acc:17.0000%
[ 2025-06-08 17:44 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 17:47 ] training: epoch: 6, loss: 11.2851, top1: 6.17%, lr: 0.001000, obj_loss: 2.9724, verb_loss:3.0868,l1_loss:0.0063, obj_acc:14.3333%, verb_acc:14.8333%
[ 2025-06-08 17:48 ] evaluating: loss: 11.4627, top1: 4.70%, best_acc: 11.30%,obj_loss: 2.9642,verb_loss:3.1927,l1_loss:0.0080,obj_acc:11.6522%, verb_acc:17.7391%
[ 2025-06-08 17:48 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 17:51 ] training: epoch: 7, loss: 10.9250, top1: 5.17%, lr: 0.001000, obj_loss: 2.8236, verb_loss:3.0495,l1_loss:0.0043, obj_acc:13.6667%, verb_acc:14.0000%
[ 2025-06-08 17:51 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 17:53 ] training: epoch: 8, loss: 10.4598, top1: 8.83%, lr: 0.001000, obj_loss: 2.6488, verb_loss:2.9696,l1_loss:0.0029, obj_acc:19.1667%, verb_acc:18.5000%
[ 2025-06-08 17:55 ] evaluating: loss: 10.4877, top1: 8.52%, best_acc: 11.30%,obj_loss: 2.6856,verb_loss:2.9736,l1_loss:0.0030,obj_acc:20.6957%, verb_acc:19.8261%
[ 2025-06-08 17:55 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 17:57 ] training: epoch: 9, loss: 10.2587, top1: 9.33%, lr: 0.001000, obj_loss: 2.5974, verb_loss:2.9230,l1_loss:0.0025, obj_acc:20.3333%, verb_acc:16.8333%
[ 2025-06-08 17:57 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 17:59 ] training: epoch: 10, loss: 10.1081, top1: 10.17%, lr: 0.001000, obj_loss: 2.5514, verb_loss:2.9026,l1_loss:0.0022, obj_acc:20.0000%, verb_acc:17.1667%
[ 2025-06-08 18:01 ] evaluating: loss: 10.3491, top1: 8.00%, best_acc: 11.30%,obj_loss: 2.6532,verb_loss:2.9145,l1_loss:0.0025,obj_acc:16.1739%, verb_acc:20.1739%
[ 2025-06-08 18:01 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 18:03 ] training: epoch: 11, loss: 10.2400, top1: 9.33%, lr: 0.001000, obj_loss: 2.5861, verb_loss:2.9140,l1_loss:0.0020, obj_acc:20.8333%, verb_acc:19.8333%
[ 2025-06-08 18:03 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 18:06 ] training: epoch: 12, loss: 9.9259, top1: 12.33%, lr: 0.001000, obj_loss: 2.5056, verb_loss:2.8409,l1_loss:0.0018, obj_acc:21.5000%, verb_acc:22.3333%
[ 2025-06-08 18:08 ] evaluating: loss: 10.3054, top1: 10.43%, best_acc: 11.30%,obj_loss: 2.5928,verb_loss:2.9453,l1_loss:0.0021,obj_acc:19.6522%, verb_acc:18.7826%
[ 2025-06-08 18:08 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 18:10 ] training: epoch: 13, loss: 10.3484, top1: 8.00%, lr: 0.001000, obj_loss: 2.6310, verb_loss:2.9367,l1_loss:0.0016, obj_acc:16.8333%, verb_acc:19.6667%
[ 2025-06-08 18:10 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 18:12 ] training: epoch: 14, loss: 10.5841, top1: 8.33%, lr: 0.001000, obj_loss: 2.7229, verb_loss:2.9780,l1_loss:0.0015, obj_acc:16.8333%, verb_acc:17.5000%
[ 2025-06-08 18:14 ] evaluating: loss: 10.2017, top1: 8.87%, best_acc: 11.30%,obj_loss: 2.5550,verb_loss:2.9406,l1_loss:0.0016,obj_acc:21.9130%, verb_acc:16.3478%
[ 2025-06-08 18:14 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 18:16 ] training: epoch: 15, loss: 10.2121, top1: 9.33%, lr: 0.001000, obj_loss: 2.5951, verb_loss:2.9044,l1_loss:0.0016, obj_acc:18.1667%, verb_acc:19.0000%
[ 2025-06-08 18:16 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 18:19 ] training: epoch: 16, loss: 10.0911, top1: 10.50%, lr: 0.001000, obj_loss: 2.5465, verb_loss:2.8852,l1_loss:0.0013, obj_acc:20.0000%, verb_acc:20.8333%
[ 2025-06-08 18:20 ] evaluating: loss: 9.9377, top1: 10.61%, best_acc: 11.30%,obj_loss: 2.5122,verb_loss:2.8349,l1_loss:0.0014,obj_acc:21.3913%, verb_acc:21.7391%
[ 2025-06-08 18:20 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 18:23 ] training: epoch: 17, loss: 9.8699, top1: 10.67%, lr: 0.001000, obj_loss: 2.4674, verb_loss:2.8476,l1_loss:0.0011, obj_acc:20.1667%, verb_acc:21.5000%
[ 2025-06-08 18:23 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 18:33 ] Loading ResNet50 weights and modifying its structure.
[ 2025-06-08 18:33 ] Model load finished: model.ISTANet.Model
[ 2025-06-08 18:33 ] Data load finished
[ 2025-06-08 18:33 ] Optimizer load finished: AdamW
[ 2025-06-08 18:33 ] action_loss_weight: 1
[ 2025-06-08 18:33 ] base_lr: 0.001
[ 2025-06-08 18:33 ] batch_size: 4
[ 2025-06-08 18:33 ] config: ./config/fpha/fpha.yaml
[ 2025-06-08 18:33 ] cuda_visible_device: 0
[ 2025-06-08 18:33 ] device: [0]
[ 2025-06-08 18:33 ] eval_interval: 2
[ 2025-06-08 18:33 ] feat_loss_weight: 1.2
[ 2025-06-08 18:33 ] feeder: feeders.feeder_fpha.Feeder
[ 2025-06-08 18:33 ] ignore_weights: []
[ 2025-06-08 18:33 ] loss: LabelSmoothingCrossEntropy
[ 2025-06-08 18:33 ] loss_args: {'smoothing': 0.15, 'temperature': 1.0}
[ 2025-06-08 18:33 ] lr_decay_rate: 0.5
[ 2025-06-08 18:33 ] model: model.ISTANet.Model
[ 2025-06-08 18:33 ] model_args: {'window_size': [20, 1, 2], 'num_frames': 120, 'num_joints': 21, 'num_persons': 2, 'num_channels': 3, 'num_classes': 45, 'num_objs': 26, 'num_verbs': 27, 'num_heads': 4, 'kernel_size': [3, 5], 'use_pes': True, 'config': [[64, 64, 16], [64, 64, 16], [64, 128, 32], [128, 128, 32], [128, 256, 64], [256, 256, 64], [256, 256, 64], [256, 256, 64]]}
[ 2025-06-08 18:33 ] nesterov: True
[ 2025-06-08 18:33 ] noun_loss_weight: 1
[ 2025-06-08 18:33 ] num_epoch: 150
[ 2025-06-08 18:33 ] num_worker: 8
[ 2025-06-08 18:33 ] optimizer: AdamW
[ 2025-06-08 18:33 ] optimizer_betas: [0.9, 0.999]
[ 2025-06-08 18:33 ] print_log: True
[ 2025-06-08 18:33 ] run_mode: train
[ 2025-06-08 18:33 ] save_epoch: 80
[ 2025-06-08 18:33 ] save_score: False
[ 2025-06-08 18:33 ] seed: 1
[ 2025-06-08 18:33 ] show_topk: [1, 5]
[ 2025-06-08 18:33 ] start_epoch: 0
[ 2025-06-08 18:33 ] step: [50, 80, 110, 130]
[ 2025-06-08 18:33 ] test_batch_size: 4
[ 2025-06-08 18:33 ] test_feeder_args: {'data_path': 'data/fpha/fpha_pth', 'split': 'val', 'debug': False, 'window_size': 120, 'p_interval': [0.95], 'vel': False, 'bone': False, 'use_rgb': True}
[ 2025-06-08 18:33 ] train_feeder_args: {'data_path': 'data/fpha/fpha_pth', 'split': 'train', 'debug': False, 'random_choose': False, 'random_shift': False, 'random_move': False, 'window_size': 120, 'normalization': False, 'random_rot': True, 'p_interval': [0.5, 1], 'vel': False, 'bone': False, 'entity_rearrangement': False, 'use_rgb': True}
[ 2025-06-08 18:33 ] verb_loss_weight: 1
[ 2025-06-08 18:33 ] warm_up_epoch: 5
[ 2025-06-08 18:33 ] weight_decay: 0.0015
[ 2025-06-08 18:33 ] weights: None
[ 2025-06-08 18:33 ] work_dir: ./exp/fpha
[ 2025-06-08 18:33 ] # Parameters: 33605890
[ 2025-06-08 18:33 ] ###***************start training***************###
[ 2025-06-08 18:33 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 18:36 ] training: epoch: 1, loss: 9.9743, top1: 3.67%, lr: 0.000200, obj_loss: 2.9951, verb_loss:3.1107,l1_loss:0.0534, obj_acc:13.3333%, verb_acc:13.6667%
[ 2025-06-08 18:36 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 18:38 ] training: epoch: 2, loss: 8.7909, top1: 7.50%, lr: 0.000400, obj_loss: 2.4246, verb_loss:2.9347,l1_loss:0.0175, obj_acc:24.5000%, verb_acc:13.8333%
[ 2025-06-08 18:40 ] evaluating: loss: 8.2686, top1: 16.35%, best_acc: 16.35%,obj_loss: 2.2715,verb_loss:2.7422,l1_loss:0.0103,obj_acc:26.4348%, verb_acc:22.0870%
[ 2025-06-08 18:40 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 18:42 ] training: epoch: 3, loss: 8.5515, top1: 10.50%, lr: 0.000600, obj_loss: 2.3735, verb_loss:2.8668,l1_loss:0.0162, obj_acc:22.6667%, verb_acc:16.1667%
[ 2025-06-08 18:42 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 18:44 ] training: epoch: 4, loss: 8.2286, top1: 14.67%, lr: 0.000800, obj_loss: 2.2170, verb_loss:2.7815,l1_loss:0.0140, obj_acc:26.0000%, verb_acc:20.0000%
[ 2025-06-08 18:46 ] evaluating: loss: 8.9298, top1: 14.43%, best_acc: 16.35%,obj_loss: 2.6025,verb_loss:2.9679,l1_loss:0.0102,obj_acc:25.5652%, verb_acc:17.3913%
[ 2025-06-08 18:46 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 18:49 ] training: epoch: 5, loss: 8.9273, top1: 8.67%, lr: 0.001000, obj_loss: 2.5083, verb_loss:2.9483,l1_loss:0.0106, obj_acc:22.1667%, verb_acc:16.0000%
[ 2025-06-08 18:49 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 18:51 ] training: epoch: 6, loss: 8.7152, top1: 7.83%, lr: 0.001000, obj_loss: 2.4143, verb_loss:2.9154,l1_loss:0.0092, obj_acc:16.8333%, verb_acc:14.5000%
[ 2025-06-08 18:53 ] evaluating: loss: 9.1075, top1: 10.61%, best_acc: 16.35%,obj_loss: 2.5300,verb_loss:3.0323,l1_loss:0.0075,obj_acc:19.1304%, verb_acc:14.0870%
[ 2025-06-08 18:53 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 18:55 ] training: epoch: 7, loss: 8.4838, top1: 9.83%, lr: 0.001000, obj_loss: 2.3313, verb_loss:2.8723,l1_loss:0.0072, obj_acc:22.5000%, verb_acc:14.3333%
[ 2025-06-08 18:55 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 18:58 ] training: epoch: 8, loss: 8.1009, top1: 11.67%, lr: 0.001000, obj_loss: 2.1543, verb_loss:2.7516,l1_loss:0.0063, obj_acc:26.1667%, verb_acc:16.5000%
[ 2025-06-08 18:59 ] evaluating: loss: 8.9934, top1: 8.17%, best_acc: 16.35%,obj_loss: 2.6352,verb_loss:2.9056,l1_loss:0.0057,obj_acc:18.9565%, verb_acc:16.0000%
[ 2025-06-08 18:59 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 19:02 ] training: epoch: 9, loss: 7.7686, top1: 14.67%, lr: 0.001000, obj_loss: 2.0446, verb_loss:2.6855,l1_loss:0.0062, obj_acc:30.5000%, verb_acc:19.1667%
[ 2025-06-08 19:02 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 19:04 ] training: epoch: 10, loss: 8.3382, top1: 15.67%, lr: 0.001000, obj_loss: 2.3600, verb_loss:2.7718,l1_loss:0.0062, obj_acc:25.3333%, verb_acc:21.1667%
[ 2025-06-08 19:06 ] evaluating: loss: 8.7730, top1: 16.52%, best_acc: 16.52%,obj_loss: 2.6166,verb_loss:2.8638,l1_loss:0.0050,obj_acc:24.8696%, verb_acc:21.3913%
[ 2025-06-08 19:06 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 19:09 ] training: epoch: 11, loss: 8.4736, top1: 12.17%, lr: 0.001000, obj_loss: 2.3903, verb_loss:2.8243,l1_loss:0.0050, obj_acc:25.1667%, verb_acc:17.3333%
[ 2025-06-08 19:09 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 19:11 ] training: epoch: 12, loss: 8.2975, top1: 14.50%, lr: 0.001000, obj_loss: 2.3410, verb_loss:2.7718,l1_loss:0.0049, obj_acc:23.8333%, verb_acc:22.0000%
[ 2025-06-08 19:13 ] evaluating: loss: 8.2883, top1: 18.43%, best_acc: 18.43%,obj_loss: 2.3600,verb_loss:2.7552,l1_loss:0.0047,obj_acc:31.6522%, verb_acc:24.8696%
[ 2025-06-08 19:13 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 19:15 ] training: epoch: 13, loss: 8.2690, top1: 13.00%, lr: 0.001000, obj_loss: 2.3348, verb_loss:2.7592,l1_loss:0.0049, obj_acc:25.5000%, verb_acc:20.3333%
[ 2025-06-08 19:15 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 19:18 ] training: epoch: 14, loss: 8.0759, top1: 15.67%, lr: 0.001000, obj_loss: 2.2542, verb_loss:2.7208,l1_loss:0.0049, obj_acc:30.3333%, verb_acc:21.0000%
[ 2025-06-08 19:20 ] evaluating: loss: 8.5588, top1: 16.35%, best_acc: 18.43%,obj_loss: 2.4858,verb_loss:2.8143,l1_loss:0.0044,obj_acc:30.2609%, verb_acc:20.8696%
[ 2025-06-08 19:20 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 19:22 ] training: epoch: 15, loss: 8.0569, top1: 13.83%, lr: 0.001000, obj_loss: 2.2705, verb_loss:2.6908,l1_loss:0.0049, obj_acc:27.0000%, verb_acc:21.3333%
[ 2025-06-08 19:22 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 19:24 ] training: epoch: 16, loss: 7.9387, top1: 15.83%, lr: 0.001000, obj_loss: 2.2193, verb_loss:2.6638,l1_loss:0.0050, obj_acc:27.5000%, verb_acc:21.5000%
[ 2025-06-08 19:26 ] evaluating: loss: 7.9382, top1: 18.61%, best_acc: 18.61%,obj_loss: 2.2478,verb_loss:2.6384,l1_loss:0.0043,obj_acc:32.3478%, verb_acc:28.3478%
[ 2025-06-08 19:26 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 19:29 ] training: epoch: 17, loss: 7.6499, top1: 19.50%, lr: 0.001000, obj_loss: 2.1100, verb_loss:2.5741,l1_loss:0.0054, obj_acc:30.6667%, verb_acc:27.1667%
[ 2025-06-08 19:29 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 19:31 ] training: epoch: 18, loss: 7.6398, top1: 19.33%, lr: 0.001000, obj_loss: 2.0846, verb_loss:2.5903,l1_loss:0.0058, obj_acc:32.0000%, verb_acc:24.1667%
[ 2025-06-08 19:33 ] evaluating: loss: 7.8525, top1: 22.78%, best_acc: 22.78%,obj_loss: 2.2071,verb_loss:2.6205,l1_loss:0.0049,obj_acc:30.7826%, verb_acc:26.9565%
[ 2025-06-08 19:33 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 19:36 ] training: epoch: 19, loss: 7.5384, top1: 22.83%, lr: 0.001000, obj_loss: 2.0648, verb_loss:2.5498,l1_loss:0.0051, obj_acc:32.6667%, verb_acc:27.1667%
[ 2025-06-08 19:36 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 19:38 ] training: epoch: 20, loss: 7.4669, top1: 18.33%, lr: 0.001000, obj_loss: 2.0380, verb_loss:2.5269,l1_loss:0.0050, obj_acc:35.0000%, verb_acc:27.3333%
[ 2025-06-08 19:40 ] evaluating: loss: 7.5862, top1: 22.96%, best_acc: 22.96%,obj_loss: 2.1089,verb_loss:2.5422,l1_loss:0.0053,obj_acc:32.5217%, verb_acc:31.1304%
[ 2025-06-08 19:40 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 19:42 ] training: epoch: 21, loss: 7.7200, top1: 18.33%, lr: 0.001000, obj_loss: 2.1461, verb_loss:2.5931,l1_loss:0.0050, obj_acc:30.0000%, verb_acc:24.8333%
[ 2025-06-08 19:42 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 19:45 ] training: epoch: 22, loss: 7.7407, top1: 22.33%, lr: 0.001000, obj_loss: 2.1539, verb_loss:2.5896,l1_loss:0.0058, obj_acc:32.1667%, verb_acc:28.3333%
[ 2025-06-08 19:47 ] evaluating: loss: 8.3601, top1: 14.43%, best_acc: 22.96%,obj_loss: 2.4187,verb_loss:2.7411,l1_loss:0.0034,obj_acc:26.2609%, verb_acc:24.6957%
[ 2025-06-08 19:47 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 19:49 ] training: epoch: 23, loss: 7.6263, top1: 19.17%, lr: 0.001000, obj_loss: 2.1075, verb_loss:2.5672,l1_loss:0.0059, obj_acc:31.8333%, verb_acc:28.5000%
[ 2025-06-08 19:49 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 19:52 ] training: epoch: 24, loss: 7.2820, top1: 21.33%, lr: 0.001000, obj_loss: 1.9599, verb_loss:2.4831,l1_loss:0.0060, obj_acc:36.1667%, verb_acc:25.8333%
[ 2025-06-08 19:53 ] evaluating: loss: 7.3480, top1: 26.26%, best_acc: 26.26%,obj_loss: 2.0388,verb_loss:2.4685,l1_loss:0.0029,obj_acc:34.0870%, verb_acc:30.4348%
[ 2025-06-08 19:53 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 19:56 ] training: epoch: 25, loss: 7.2184, top1: 26.00%, lr: 0.001000, obj_loss: 1.9391, verb_loss:2.4539,l1_loss:0.0054, obj_acc:35.8333%, verb_acc:31.5000%
[ 2025-06-08 19:56 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 19:58 ] training: epoch: 26, loss: 7.1487, top1: 27.67%, lr: 0.001000, obj_loss: 1.8855, verb_loss:2.4619,l1_loss:0.0047, obj_acc:40.8333%, verb_acc:29.5000%
[ 2025-06-08 20:00 ] evaluating: loss: 7.4431, top1: 25.91%, best_acc: 26.26%,obj_loss: 2.0063,verb_loss:2.5411,l1_loss:0.0039,obj_acc:38.9565%, verb_acc:28.5217%
[ 2025-06-08 20:00 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 20:02 ] training: epoch: 27, loss: 6.9704, top1: 26.83%, lr: 0.001000, obj_loss: 1.8180, verb_loss:2.4125,l1_loss:0.0056, obj_acc:41.5000%, verb_acc:31.5000%
[ 2025-06-08 20:02 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 20:05 ] Loading ResNet50 weights and modifying its structure.
[ 2025-06-08 20:05 ] Model load finished: model.ISTANet.Model
[ 2025-06-08 20:05 ] Data load finished
[ 2025-06-08 20:05 ] Optimizer load finished: AdamW
[ 2025-06-08 20:05 ] action_loss_weight: 1
[ 2025-06-08 20:05 ] base_lr: 0.001
[ 2025-06-08 20:05 ] batch_size: 4
[ 2025-06-08 20:05 ] config: ./config/fpha/fpha.yaml
[ 2025-06-08 20:05 ] cuda_visible_device: 0
[ 2025-06-08 20:05 ] device: [0]
[ 2025-06-08 20:05 ] eval_interval: 2
[ 2025-06-08 20:05 ] feat_loss_weight: 1.2
[ 2025-06-08 20:05 ] feeder: feeders.feeder_fpha.Feeder
[ 2025-06-08 20:05 ] ignore_weights: []
[ 2025-06-08 20:05 ] loss: LabelSmoothingCrossEntropy
[ 2025-06-08 20:05 ] loss_args: {'smoothing': 0.15, 'temperature': 1.0}
[ 2025-06-08 20:05 ] lr_decay_rate: 0.5
[ 2025-06-08 20:05 ] model: model.ISTANet.Model
[ 2025-06-08 20:05 ] model_args: {'window_size': [20, 1, 2], 'num_frames': 120, 'num_joints': 21, 'num_persons': 2, 'num_channels': 3, 'num_classes': 45, 'num_objs': 26, 'num_verbs': 27, 'num_heads': 4, 'kernel_size': [3, 5], 'use_pes': True, 'config': [[64, 64, 16], [64, 64, 16], [64, 128, 32], [128, 128, 32], [128, 256, 64], [256, 256, 64], [256, 256, 64], [256, 256, 64]]}
[ 2025-06-08 20:05 ] nesterov: True
[ 2025-06-08 20:05 ] noun_loss_weight: 1
[ 2025-06-08 20:05 ] num_epoch: 150
[ 2025-06-08 20:05 ] num_worker: 8
[ 2025-06-08 20:05 ] optimizer: AdamW
[ 2025-06-08 20:05 ] optimizer_betas: [0.9, 0.999]
[ 2025-06-08 20:05 ] print_log: True
[ 2025-06-08 20:05 ] run_mode: train
[ 2025-06-08 20:05 ] save_epoch: 80
[ 2025-06-08 20:05 ] save_score: False
[ 2025-06-08 20:05 ] seed: 1
[ 2025-06-08 20:05 ] show_topk: [1, 5]
[ 2025-06-08 20:05 ] start_epoch: 0
[ 2025-06-08 20:05 ] step: [50, 80, 110, 130]
[ 2025-06-08 20:05 ] test_batch_size: 4
[ 2025-06-08 20:05 ] test_feeder_args: {'data_path': 'data/fpha/fpha_pth', 'split': 'val', 'debug': False, 'window_size': 120, 'p_interval': [0.95], 'vel': False, 'bone': False, 'use_rgb': True}
[ 2025-06-08 20:05 ] train_feeder_args: {'data_path': 'data/fpha/fpha_pth', 'split': 'train', 'debug': False, 'random_choose': False, 'random_shift': False, 'random_move': False, 'window_size': 120, 'normalization': False, 'random_rot': True, 'p_interval': [0.5, 1], 'vel': False, 'bone': False, 'entity_rearrangement': False, 'use_rgb': True}
[ 2025-06-08 20:05 ] verb_loss_weight: 0
[ 2025-06-08 20:05 ] warm_up_epoch: 5
[ 2025-06-08 20:05 ] weight_decay: 0.0015
[ 2025-06-08 20:05 ] weights: None
[ 2025-06-08 20:05 ] work_dir: ./exp/fpha
[ 2025-06-08 20:05 ] # Parameters: 33605890
[ 2025-06-08 20:05 ] ###***************start training***************###
[ 2025-06-08 20:05 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 20:08 ] training: epoch: 1, loss: 6.8286, top1: 5.00%, lr: 0.000200, obj_loss: 2.9686, verb_loss:3.3920,l1_loss:0.0531, obj_acc:14.3333%, verb_acc:1.0000%
[ 2025-06-08 20:08 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 20:10 ] training: epoch: 2, loss: 5.8256, top1: 9.83%, lr: 0.000400, obj_loss: 2.4060, verb_loss:3.4287,l1_loss:0.0172, obj_acc:25.3333%, verb_acc:1.6667%
[ 2025-06-08 20:12 ] evaluating: loss: 5.4605, top1: 17.22%, best_acc: 17.22%,obj_loss: 2.2014,verb_loss:3.4665,l1_loss:0.0094,obj_acc:32.5217%, verb_acc:0.1739%
[ 2025-06-08 20:12 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 20:15 ] training: epoch: 3, loss: 5.9689, top1: 8.83%, lr: 0.000600, obj_loss: 2.4925, verb_loss:3.4254,l1_loss:0.0163, obj_acc:22.3333%, verb_acc:1.8333%
[ 2025-06-08 20:15 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 20:17 ] training: epoch: 4, loss: 5.6747, top1: 10.83%, lr: 0.000800, obj_loss: 2.3345, verb_loss:3.3967,l1_loss:0.0129, obj_acc:23.0000%, verb_acc:1.3333%
[ 2025-06-08 20:19 ] evaluating: loss: 6.5973, top1: 10.96%, best_acc: 17.22%,obj_loss: 2.9261,verb_loss:3.4246,l1_loss:0.0083,obj_acc:24.1739%, verb_acc:3.8261%
[ 2025-06-08 20:19 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 20:21 ] training: epoch: 5, loss: 6.8108, top1: 5.67%, lr: 0.001000, obj_loss: 3.0003, verb_loss:3.3495,l1_loss:0.0099, obj_acc:16.3333%, verb_acc:2.5000%
[ 2025-06-08 20:21 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 20:24 ] training: epoch: 6, loss: 6.7692, top1: 2.67%, lr: 0.001000, obj_loss: 2.9947, verb_loss:3.3332,l1_loss:0.0092, obj_acc:11.0000%, verb_acc:2.1667%
[ 2025-06-08 20:26 ] evaluating: loss: 6.6598, top1: 5.74%, best_acc: 17.22%,obj_loss: 2.9339,verb_loss:3.3920,l1_loss:0.0095,obj_acc:14.6087%, verb_acc:2.0870%
[ 2025-06-08 20:26 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 20:28 ] training: epoch: 7, loss: 6.5063, top1: 4.67%, lr: 0.001000, obj_loss: 2.8508, verb_loss:3.3727,l1_loss:0.0086, obj_acc:12.6667%, verb_acc:2.3333%
[ 2025-06-08 20:28 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 20:33 ] Loading ResNet50 weights and modifying its structure.
[ 2025-06-08 20:33 ] Model load finished: model.ISTANet.Model
[ 2025-06-08 20:33 ] Data load finished
[ 2025-06-08 20:33 ] Optimizer load finished: AdamW
[ 2025-06-08 20:33 ] action_loss_weight: 1
[ 2025-06-08 20:33 ] base_lr: 0.001
[ 2025-06-08 20:33 ] batch_size: 4
[ 2025-06-08 20:33 ] config: ./config/fpha/fpha.yaml
[ 2025-06-08 20:33 ] cuda_visible_device: 0
[ 2025-06-08 20:33 ] device: [0]
[ 2025-06-08 20:33 ] eval_interval: 2
[ 2025-06-08 20:33 ] feat_loss_weight: 1.2
[ 2025-06-08 20:33 ] feeder: feeders.feeder_fpha.Feeder
[ 2025-06-08 20:33 ] ignore_weights: []
[ 2025-06-08 20:33 ] loss: LabelSmoothingCrossEntropy
[ 2025-06-08 20:33 ] loss_args: {'smoothing': 0.15, 'temperature': 1.0}
[ 2025-06-08 20:33 ] lr_decay_rate: 0.5
[ 2025-06-08 20:33 ] model: model.ISTANet.Model
[ 2025-06-08 20:33 ] model_args: {'window_size': [20, 1, 2], 'num_frames': 120, 'num_joints': 21, 'num_persons': 2, 'num_channels': 3, 'num_classes': 45, 'num_objs': 26, 'num_verbs': 27, 'num_heads': 4, 'kernel_size': [3, 5], 'use_pes': True, 'config': [[64, 64, 16], [64, 64, 16], [64, 128, 32], [128, 128, 32], [128, 256, 64], [256, 256, 64], [256, 256, 64], [256, 256, 64]]}
[ 2025-06-08 20:33 ] nesterov: True
[ 2025-06-08 20:33 ] noun_loss_weight: 1
[ 2025-06-08 20:33 ] num_epoch: 150
[ 2025-06-08 20:33 ] num_worker: 8
[ 2025-06-08 20:33 ] optimizer: AdamW
[ 2025-06-08 20:33 ] optimizer_betas: [0.9, 0.999]
[ 2025-06-08 20:33 ] print_log: True
[ 2025-06-08 20:33 ] run_mode: train
[ 2025-06-08 20:33 ] save_epoch: 80
[ 2025-06-08 20:33 ] save_score: False
[ 2025-06-08 20:33 ] seed: 1
[ 2025-06-08 20:33 ] show_topk: [1, 5]
[ 2025-06-08 20:33 ] start_epoch: 0
[ 2025-06-08 20:33 ] step: [50, 80, 110, 130]
[ 2025-06-08 20:33 ] test_batch_size: 4
[ 2025-06-08 20:33 ] test_feeder_args: {'data_path': 'data/fpha/fpha_pth', 'split': 'val', 'debug': False, 'window_size': 120, 'p_interval': [0.95], 'vel': False, 'bone': False, 'use_rgb': True}
[ 2025-06-08 20:33 ] train_feeder_args: {'data_path': 'data/fpha/fpha_pth', 'split': 'train', 'debug': False, 'random_choose': False, 'random_shift': False, 'random_move': False, 'window_size': 120, 'normalization': False, 'random_rot': True, 'p_interval': [0.5, 1], 'vel': False, 'bone': False, 'entity_rearrangement': False, 'use_rgb': True}
[ 2025-06-08 20:33 ] verb_loss_weight: 0
[ 2025-06-08 20:33 ] warm_up_epoch: 5
[ 2025-06-08 20:33 ] weight_decay: 0.0015
[ 2025-06-08 20:33 ] weights: None
[ 2025-06-08 20:33 ] work_dir: ./exp/fpha
[ 2025-06-08 20:33 ] # Parameters: 33592039
[ 2025-06-08 20:33 ] ###***************start training***************###
[ 2025-06-08 20:33 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 20:34 ] Loading ResNet50 weights and modifying its structure.
[ 2025-06-08 20:34 ] Model load finished: model.ISTANet.Model
[ 2025-06-08 20:34 ] Data load finished
[ 2025-06-08 20:34 ] Optimizer load finished: AdamW
[ 2025-06-08 20:34 ] action_loss_weight: 1
[ 2025-06-08 20:34 ] base_lr: 0.001
[ 2025-06-08 20:34 ] batch_size: 4
[ 2025-06-08 20:34 ] config: ./config/fpha/fpha.yaml
[ 2025-06-08 20:34 ] cuda_visible_device: 0
[ 2025-06-08 20:34 ] device: [0]
[ 2025-06-08 20:34 ] eval_interval: 2
[ 2025-06-08 20:34 ] feat_loss_weight: 1.2
[ 2025-06-08 20:34 ] feeder: feeders.feeder_fpha.Feeder
[ 2025-06-08 20:34 ] ignore_weights: []
[ 2025-06-08 20:34 ] loss: LabelSmoothingCrossEntropy
[ 2025-06-08 20:34 ] loss_args: {'smoothing': 0.15, 'temperature': 1.0}
[ 2025-06-08 20:34 ] lr_decay_rate: 0.5
[ 2025-06-08 20:34 ] model: model.ISTANet.Model
[ 2025-06-08 20:34 ] model_args: {'window_size': [20, 1, 2], 'num_frames': 120, 'num_joints': 21, 'num_persons': 2, 'num_channels': 3, 'num_classes': 45, 'num_objs': 26, 'num_verbs': 27, 'num_heads': 4, 'kernel_size': [3, 5], 'use_pes': True, 'config': [[64, 64, 16], [64, 64, 16], [64, 128, 32], [128, 128, 32], [128, 256, 64], [256, 256, 64], [256, 256, 64], [256, 256, 64]]}
[ 2025-06-08 20:34 ] nesterov: True
[ 2025-06-08 20:34 ] noun_loss_weight: 1
[ 2025-06-08 20:34 ] num_epoch: 150
[ 2025-06-08 20:34 ] num_worker: 8
[ 2025-06-08 20:34 ] optimizer: AdamW
[ 2025-06-08 20:34 ] optimizer_betas: [0.9, 0.999]
[ 2025-06-08 20:34 ] print_log: True
[ 2025-06-08 20:34 ] run_mode: train
[ 2025-06-08 20:34 ] save_epoch: 80
[ 2025-06-08 20:34 ] save_score: False
[ 2025-06-08 20:34 ] seed: 1
[ 2025-06-08 20:34 ] show_topk: [1, 5]
[ 2025-06-08 20:34 ] start_epoch: 0
[ 2025-06-08 20:34 ] step: [50, 80, 110, 130]
[ 2025-06-08 20:34 ] test_batch_size: 4
[ 2025-06-08 20:34 ] test_feeder_args: {'data_path': 'data/fpha/fpha_pth', 'split': 'val', 'debug': False, 'window_size': 120, 'p_interval': [0.95], 'vel': False, 'bone': False, 'use_rgb': True}
[ 2025-06-08 20:34 ] train_feeder_args: {'data_path': 'data/fpha/fpha_pth', 'split': 'train', 'debug': False, 'random_choose': False, 'random_shift': False, 'random_move': False, 'window_size': 120, 'normalization': False, 'random_rot': True, 'p_interval': [0.5, 1], 'vel': False, 'bone': False, 'entity_rearrangement': False, 'use_rgb': True}
[ 2025-06-08 20:34 ] verb_loss_weight: 0
[ 2025-06-08 20:34 ] warm_up_epoch: 5
[ 2025-06-08 20:34 ] weight_decay: 0.0015
[ 2025-06-08 20:34 ] weights: None
[ 2025-06-08 20:34 ] work_dir: ./exp/fpha
[ 2025-06-08 20:34 ] # Parameters: 33592039
[ 2025-06-08 20:34 ] ###***************start training***************###
[ 2025-06-08 20:34 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 20:37 ] Loading ResNet50 weights and modifying its structure.
[ 2025-06-08 20:37 ] Model load finished: model.ISTANet.Model
[ 2025-06-08 20:37 ] Data load finished
[ 2025-06-08 20:37 ] Optimizer load finished: AdamW
[ 2025-06-08 20:37 ] action_loss_weight: 1
[ 2025-06-08 20:37 ] base_lr: 0.001
[ 2025-06-08 20:37 ] batch_size: 4
[ 2025-06-08 20:37 ] config: ./config/fpha/fpha.yaml
[ 2025-06-08 20:37 ] cuda_visible_device: 0
[ 2025-06-08 20:37 ] device: [0]
[ 2025-06-08 20:37 ] eval_interval: 2
[ 2025-06-08 20:37 ] feat_loss_weight: 1.2
[ 2025-06-08 20:37 ] feeder: feeders.feeder_fpha.Feeder
[ 2025-06-08 20:37 ] ignore_weights: []
[ 2025-06-08 20:37 ] loss: LabelSmoothingCrossEntropy
[ 2025-06-08 20:37 ] loss_args: {'smoothing': 0.15, 'temperature': 1.0}
[ 2025-06-08 20:37 ] lr_decay_rate: 0.5
[ 2025-06-08 20:37 ] model: model.ISTANet.Model
[ 2025-06-08 20:37 ] model_args: {'window_size': [20, 1, 2], 'num_frames': 120, 'num_joints': 21, 'num_persons': 2, 'num_channels': 3, 'num_classes': 45, 'num_objs': 26, 'num_verbs': 27, 'num_heads': 4, 'kernel_size': [3, 5], 'use_pes': True, 'config': [[64, 64, 16], [64, 64, 16], [64, 128, 32], [128, 128, 32], [128, 256, 64], [256, 256, 64], [256, 256, 64], [256, 256, 64]]}
[ 2025-06-08 20:37 ] nesterov: True
[ 2025-06-08 20:37 ] noun_loss_weight: 1
[ 2025-06-08 20:37 ] num_epoch: 150
[ 2025-06-08 20:37 ] num_worker: 8
[ 2025-06-08 20:37 ] optimizer: AdamW
[ 2025-06-08 20:37 ] optimizer_betas: [0.9, 0.999]
[ 2025-06-08 20:37 ] print_log: True
[ 2025-06-08 20:37 ] run_mode: train
[ 2025-06-08 20:37 ] save_epoch: 80
[ 2025-06-08 20:37 ] save_score: False
[ 2025-06-08 20:37 ] seed: 1
[ 2025-06-08 20:37 ] show_topk: [1, 5]
[ 2025-06-08 20:37 ] start_epoch: 0
[ 2025-06-08 20:37 ] step: [50, 80, 110, 130]
[ 2025-06-08 20:37 ] test_batch_size: 4
[ 2025-06-08 20:37 ] test_feeder_args: {'data_path': 'data/fpha/fpha_pth', 'split': 'val', 'debug': False, 'window_size': 120, 'p_interval': [0.95], 'vel': False, 'bone': False, 'use_rgb': True}
[ 2025-06-08 20:37 ] train_feeder_args: {'data_path': 'data/fpha/fpha_pth', 'split': 'train', 'debug': False, 'random_choose': False, 'random_shift': False, 'random_move': False, 'window_size': 120, 'normalization': False, 'random_rot': True, 'p_interval': [0.5, 1], 'vel': False, 'bone': False, 'entity_rearrangement': False, 'use_rgb': True}
[ 2025-06-08 20:37 ] verb_loss_weight: 0
[ 2025-06-08 20:37 ] warm_up_epoch: 5
[ 2025-06-08 20:37 ] weight_decay: 0.0015
[ 2025-06-08 20:37 ] weights: None
[ 2025-06-08 20:37 ] work_dir: ./exp/fpha
[ 2025-06-08 20:37 ] # Parameters: 33592039
[ 2025-06-08 20:37 ] ###***************start training***************###
[ 2025-06-08 20:37 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 20:38 ] Loading ResNet50 weights and modifying its structure.
[ 2025-06-08 20:38 ] Model load finished: model.ISTANet.Model
[ 2025-06-08 20:38 ] Data load finished
[ 2025-06-08 20:38 ] Optimizer load finished: AdamW
[ 2025-06-08 20:38 ] action_loss_weight: 1
[ 2025-06-08 20:38 ] base_lr: 0.001
[ 2025-06-08 20:38 ] batch_size: 4
[ 2025-06-08 20:38 ] config: ./config/fpha/fpha.yaml
[ 2025-06-08 20:38 ] cuda_visible_device: 0
[ 2025-06-08 20:38 ] device: [0]
[ 2025-06-08 20:38 ] eval_interval: 2
[ 2025-06-08 20:38 ] feat_loss_weight: 1.2
[ 2025-06-08 20:38 ] feeder: feeders.feeder_fpha.Feeder
[ 2025-06-08 20:38 ] ignore_weights: []
[ 2025-06-08 20:38 ] loss: LabelSmoothingCrossEntropy
[ 2025-06-08 20:38 ] loss_args: {'smoothing': 0.15, 'temperature': 1.0}
[ 2025-06-08 20:38 ] lr_decay_rate: 0.5
[ 2025-06-08 20:38 ] model: model.ISTANet.Model
[ 2025-06-08 20:38 ] model_args: {'window_size': [20, 1, 2], 'num_frames': 120, 'num_joints': 21, 'num_persons': 2, 'num_channels': 3, 'num_classes': 45, 'num_objs': 26, 'num_verbs': 27, 'num_heads': 4, 'kernel_size': [3, 5], 'use_pes': True, 'config': [[64, 64, 16], [64, 64, 16], [64, 128, 32], [128, 128, 32], [128, 256, 64], [256, 256, 64], [256, 256, 64], [256, 256, 64]]}
[ 2025-06-08 20:38 ] nesterov: True
[ 2025-06-08 20:38 ] noun_loss_weight: 1
[ 2025-06-08 20:38 ] num_epoch: 150
[ 2025-06-08 20:38 ] num_worker: 8
[ 2025-06-08 20:38 ] optimizer: AdamW
[ 2025-06-08 20:38 ] optimizer_betas: [0.9, 0.999]
[ 2025-06-08 20:38 ] print_log: True
[ 2025-06-08 20:38 ] run_mode: train
[ 2025-06-08 20:38 ] save_epoch: 80
[ 2025-06-08 20:38 ] save_score: False
[ 2025-06-08 20:38 ] seed: 1
[ 2025-06-08 20:38 ] show_topk: [1, 5]
[ 2025-06-08 20:38 ] start_epoch: 0
[ 2025-06-08 20:38 ] step: [50, 80, 110, 130]
[ 2025-06-08 20:38 ] test_batch_size: 4
[ 2025-06-08 20:38 ] test_feeder_args: {'data_path': 'data/fpha/fpha_pth', 'split': 'val', 'debug': False, 'window_size': 120, 'p_interval': [0.95], 'vel': False, 'bone': False, 'use_rgb': True}
[ 2025-06-08 20:38 ] train_feeder_args: {'data_path': 'data/fpha/fpha_pth', 'split': 'train', 'debug': False, 'random_choose': False, 'random_shift': False, 'random_move': False, 'window_size': 120, 'normalization': False, 'random_rot': True, 'p_interval': [0.5, 1], 'vel': False, 'bone': False, 'entity_rearrangement': False, 'use_rgb': True}
[ 2025-06-08 20:38 ] verb_loss_weight: 0
[ 2025-06-08 20:38 ] warm_up_epoch: 5
[ 2025-06-08 20:38 ] weight_decay: 0.0015
[ 2025-06-08 20:38 ] weights: None
[ 2025-06-08 20:38 ] work_dir: ./exp/fpha
[ 2025-06-08 20:38 ] # Parameters: 33592039
[ 2025-06-08 20:38 ] ###***************start training***************###
[ 2025-06-08 20:38 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 20:40 ] training: epoch: 1, loss: 6.8618, top1: 3.67%, lr: 0.000200, obj_loss: 2.9689,l1_loss:0.0533, obj_acc:16.1667%
[ 2025-06-08 20:40 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 20:42 ] training: epoch: 2, loss: 5.8699, top1: 9.17%, lr: 0.000400, obj_loss: 2.4254,l1_loss:0.0173, obj_acc:24.8333%
[ 2025-06-08 20:44 ] evaluating: loss: 6.0352, top1: 9.57%, best_acc: 9.57%,obj_loss: 2.5072,l1_loss:0.0109,obj_acc:28.1739%
[ 2025-06-08 20:44 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 20:47 ] training: epoch: 3, loss: 5.7118, top1: 7.67%, lr: 0.000600, obj_loss: 2.3167,l1_loss:0.0163, obj_acc:23.1667%
[ 2025-06-08 20:47 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 20:49 ] training: epoch: 4, loss: 6.2033, top1: 10.00%, lr: 0.000800, obj_loss: 2.6681,l1_loss:0.0132, obj_acc:21.5000%
[ 2025-06-08 20:51 ] evaluating: loss: 7.1740, top1: 5.04%, best_acc: 9.57%,obj_loss: 3.2495,l1_loss:0.0074,obj_acc:15.4783%
[ 2025-06-08 20:51 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-08 20:53 ] training: epoch: 5, loss: 6.5424, top1: 3.83%, lr: 0.001000, obj_loss: 2.8323,l1_loss:0.0095, obj_acc:15.3333%
[ 2025-06-08 20:53 ] adjust learning rate, using warm up, epoch: 5
[ 2025-06-14 16:12 ] Load weights from D:/Downloads/fpha final 93.57/fpha.pth
[ 2025-06-14 16:13 ] Load weights from D:/Downloads/fpha final 93.57/fpha.pt
[ 2025-06-14 16:13 ] Model load finished: model.ISTANet.Model
[ 2025-06-14 16:13 ] Data load finished
