File size: 131,927 Bytes
f1c1609 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 | backup evironment completed !
Loading pth from /mnt/data/pjlab-3090-sport/wuhao/logs/dibs/howto-anet_anet_ori_pbox(similarity_op_order_v2)_CLIP/similarity_op_order_v2_topf30_iter3_r1_th1_refine_aug(8,0.02)_top3_2stage_ins_cap_topk_mil_coef0_noFocal
******************** All args: *************************************************
align_contiguous = False
align_drop_z = 0
align_keep_percentile = 0.1
align_many_to_one = False
align_one_to_many = False
align_top_band_size = 0
att_hid_size = 512
aux_loss = True
backbone = None
base_cfg_path = cfgs_base/howto/howto-anet_anet_ori_(sim_op_order_v2)_CLIP_refine.yml
basic_ss_prob = 0
batch_size = 1
batch_size_for_eval = 1
bbox_loss_coef = 0
beta = 1
cap_dec_n_points = 4
cap_nheads = 1
cap_num_feature_levels = 4
cap_prob_clip = False
caption_cost_type = loss
caption_decoder_type = standard
caption_loss_coef = 2
cfg_path = cfgs/howto-anet_anet_clip_topk30_r1_iter3_th1_refine_aug(8,0.02)_top3_2stage_inscap.yml
cl_schedule_time = [0, 2]
cl_schedule_val = [0, 0.1]
clip_context_dim = 512
cls_loss_coef = 2
contrastive_hidden_size = 128
contrastive_loss_start_coef = 0.0
contrastive_loss_temperature = 0.1
cost_alpha = 0.25
cost_gamma = 2
count_loss_coef = 0.5
criteria_for_best_ckpt = overall
current_lr = 5e-05
data_norm = 0
data_rescale = 1
debug = False
dec_layers = 2
dec_n_points = 4
device = cuda
dict_file = data/howto/vocabulary_howto_rate2_anet.json
dict_file_val = data/howto/vocabulary_howto_rate2_anet.json
dilation = False
disable_contrastive_projection = 1
disable_cudnn = 0
disable_mid_caption_heads = False
disable_rematch = False
disable_tqdm = False
drop_prob = 0.5
ec_alpha = 1.0
enable_bg_for_cl = True
enable_contrastive = False
enable_cross_video_cl = True
enable_e2t_cl = True
enc_layers = 2
enc_n_points = 4
eos_coef = 0.1
epoch = 20
eval_proposal_file = data/generated_proposals/dbg_trainval_top100.json
event_context_dim = None
feature_dim = 768
feature_sample_rate = 1
fix_xcw = 1
focal_alpha = 0.25
focal_gamma = 2.0
focal_mil = False
frame_embedding_num = 100
ft_gt_percent = 1.0
giou_loss_coef = 4
gpu_id = []
grad_clip = 100.0
gt_file_for_auc = data/anet/captiondata/val_all.json
gt_file_for_eval = ['data/anet/captiondata/val_1.json', 'data/anet/captiondata/val_2.json']
gt_file_for_para_eval = ['data/anet/captiondata/para/anet_entities_val_1_para.json', 'data/anet/captiondata/para/anet_entities_val_2_para.json']
gt_proposal_sample_num = 20
hidden_dim = 512
hidden_dropout_prob = 0.5
huggingface_cache_dir = .cache
id = seq2-ft(mix)-gt_percent-1.0
id_ori =
input_encoding_size = 512
invalid_video_json = []
iteration = 3
layer_norm_eps = 1e-12
learning_rate_decay_every = 3
learning_rate_decay_rate = 0.5
learning_rate_decay_start = 8
lloss_beta = 1
lloss_cross_entropy = 0
lloss_focal_loss = 0
lloss_gau_mask = 1
lr = 5e-05
lr_backbone = 2e-05
lr_backbone_names = ['None']
lr_linear_proj_mult = 0.1
lr_linear_proj_names = ['reference_points', 'sampling_offsets']
lr_proj = 0
map = True
matcher_type = default
max_caption_len = 50
max_eseq_length = 10
max_pos_num = 500
max_text_input_len = 32
merge_criterion = ins_cap_topk
merge_k_boxes = 3
merge_mode = weighted_sum
mil_loss_coef = 0
min_epoch_when_save = -1
nheads = 8
norm_ins_score = sigmoid
nthreads = 4
num_classes = 1
num_feature_levels = 4
num_layers = 1
num_neg_box = 10
num_queries = 100
optimizer_type = adam
position_embedding = sine
position_embedding_scale = 6.283185307179586
pre_percent = 1.0
pretrain = None
pretrain_path =
pretrained_language_model = CLIP
prior_anchor_duration_init = True
prior_manner = all
pseudo_box_aug = False
pseudo_box_aug_mode = random_range
pseudo_box_aug_num = 8
pseudo_box_aug_ratio = 0.02
pseudo_box_type = similarity_op_order_v2
random_anchor_init = True
random_seed = False
ref_rank_loss_coef = 0.0
refine_pseudo_box = False
refine_pseudo_stage_num = 2
rnn_size = 512
sample_method = nearest
save_all_checkpoint = 0
save_checkpoint_every = 1
save_dir = /mnt/data/pjlab-3090-sport/wuhao/logs/dibs
scheduled_sampling_increase_every = 2
scheduled_sampling_increase_prob = 0.05
scheduled_sampling_max_prob = 0.25
scheduled_sampling_start = -1
seed = 777
self_iou_loss_coef = 0.0
set_cost_bbox = 0
set_cost_caption = 0
set_cost_cl = 0.0
set_cost_class = 2
set_cost_giou = 4
set_cost_sim = 1.0
share_caption_head = 1
soft_attention = 1
start_from =
start_from_mode = last
start_refine_epoch = -1
statistic_mode = mode
text_encoder_learning_strategy = frozen
text_feature_folder = ['/mnt/data/Gvlab/wuhao/features/howto100m/clip/text_proj', '/mnt/data/Gvlab/wuhao/features/anet/CLIP_feature/text/']
text_feature_folder_val = ['/mnt/data/Gvlab/wuhao/features/anet/CLIP_feature/text/']
text_hidden_dim = 768
top_frames = 30
train_caption_file = ['data/howto/captiondata/howto100m_train.json', 'data/anet/captiondata/train_modified.json']
train_proposal_file = data/generated_proposals/dbg_trainval_top100.json
train_proposal_sample_num = 30
train_proposal_type = gt
training_scheme = all
transformer_dropout_prob = 0.1
transformer_ff_dim = 512
transformer_input_type = queries
use_additional_cap_layer = False
use_additional_score_layer = False
use_anchor = 0
use_neg_pseudo_box = False
use_pseudo_box = False
use_query_box_for_refine = 0
val_caption_file = data/anet/captiondata/val_1.json
visual_feature_folder = ['/mnt/data/Gvlab/wuhao/features/howto100m/clip/visual', '/mnt/data/Gvlab/wuhao/features/anet/CLIP_feature/visual/']
visual_feature_folder_val = ['/mnt/data/Gvlab/wuhao/features/anet/CLIP_feature/visual/']
visual_feature_type = ['CLIP']
vocab_size = 16221
vocab_size_val = 16221
weight_decay = 0.0001
weighted_mil_loss = False
width_ratio = 1
width_th = 1
window_size = 2
with_box_refine = 1
wordRNN_input_feats_type = C
******************** Model structure: ******************************************
PDVC(
(base_encoder): BaseEncoder(
(pos_embed): PositionEmbeddingSine(
(duration_embed_layer): Linear(in_features=256, out_features=256, bias=True)
)
(input_proj): ModuleList(
(0): Sequential(
(0): Conv1d(768, 512, kernel_size=(1,), stride=(1,))
(1): GroupNorm(32, 512, eps=1e-05, affine=True)
)
(1): Sequential(
(0): Conv1d(768, 512, kernel_size=(3,), stride=(2,), padding=(1,))
(1): GroupNorm(32, 512, eps=1e-05, affine=True)
)
(2): Sequential(
(0): Conv1d(512, 512, kernel_size=(3,), stride=(2,), padding=(1,))
(1): GroupNorm(32, 512, eps=1e-05, affine=True)
)
(3): Sequential(
(0): Conv1d(512, 512, kernel_size=(3,), stride=(2,), padding=(1,))
(1): GroupNorm(32, 512, eps=1e-05, affine=True)
)
)
)
(transformer): DeformableTransformer(
(encoder): DeformableTransformerEncoder(
(layers): ModuleList(
(0): DeformableTransformerEncoderLayer(
(self_attn): MSDeformAttn(
(sampling_offsets): Linear(in_features=512, out_features=128, bias=True)
(attention_weights): Linear(in_features=512, out_features=128, bias=True)
(value_proj): Linear(in_features=512, out_features=512, bias=True)
(output_proj): Linear(in_features=512, out_features=512, bias=True)
)
(dropout1): Dropout(p=0.1, inplace=False)
(norm1): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
(linear1): Linear(in_features=512, out_features=512, bias=True)
(dropout2): Dropout(p=0.1, inplace=False)
(linear2): Linear(in_features=512, out_features=512, bias=True)
(dropout3): Dropout(p=0.1, inplace=False)
(norm2): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
)
(1): DeformableTransformerEncoderLayer(
(self_attn): MSDeformAttn(
(sampling_offsets): Linear(in_features=512, out_features=128, bias=True)
(attention_weights): Linear(in_features=512, out_features=128, bias=True)
(value_proj): Linear(in_features=512, out_features=512, bias=True)
(output_proj): Linear(in_features=512, out_features=512, bias=True)
)
(dropout1): Dropout(p=0.1, inplace=False)
(norm1): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
(linear1): Linear(in_features=512, out_features=512, bias=True)
(dropout2): Dropout(p=0.1, inplace=False)
(linear2): Linear(in_features=512, out_features=512, bias=True)
(dropout3): Dropout(p=0.1, inplace=False)
(norm2): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
)
)
)
(decoder): DeformableTransformerDecoder(
(layers): ModuleList(
(0): DeformableTransformerDecoderLayer(
(cross_attn): MSDeformAttn(
(sampling_offsets): Linear(in_features=512, out_features=128, bias=True)
(attention_weights): Linear(in_features=512, out_features=128, bias=True)
(value_proj): Linear(in_features=512, out_features=512, bias=True)
(output_proj): Linear(in_features=512, out_features=512, bias=True)
)
(dropout1): Dropout(p=0.1, inplace=False)
(norm1): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
(self_attn): MultiheadAttention(
(out_proj): NonDynamicallyQuantizableLinear(in_features=512, out_features=512, bias=True)
)
(dropout2): Dropout(p=0.1, inplace=False)
(norm2): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
(linear1): Linear(in_features=512, out_features=512, bias=True)
(dropout3): Dropout(p=0.1, inplace=False)
(linear2): Linear(in_features=512, out_features=512, bias=True)
(dropout4): Dropout(p=0.1, inplace=False)
(norm3): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
)
(1): DeformableTransformerDecoderLayer(
(cross_attn): MSDeformAttn(
(sampling_offsets): Linear(in_features=512, out_features=128, bias=True)
(attention_weights): Linear(in_features=512, out_features=128, bias=True)
(value_proj): Linear(in_features=512, out_features=512, bias=True)
(output_proj): Linear(in_features=512, out_features=512, bias=True)
)
(dropout1): Dropout(p=0.1, inplace=False)
(norm1): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
(self_attn): MultiheadAttention(
(out_proj): NonDynamicallyQuantizableLinear(in_features=512, out_features=512, bias=True)
)
(dropout2): Dropout(p=0.1, inplace=False)
(norm2): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
(linear1): Linear(in_features=512, out_features=512, bias=True)
(dropout3): Dropout(p=0.1, inplace=False)
(linear2): Linear(in_features=512, out_features=512, bias=True)
(dropout4): Dropout(p=0.1, inplace=False)
(norm3): LayerNorm((512,), eps=1e-05, elementwise_affine=True)
)
)
(bbox_head): ModuleList(
(0): MLP(
(layers): ModuleList(
(0): Linear(in_features=512, out_features=512, bias=True)
(1): Linear(in_features=512, out_features=512, bias=True)
(2): Linear(in_features=512, out_features=2, bias=True)
)
)
(1): MLP(
(layers): ModuleList(
(0): Linear(in_features=512, out_features=512, bias=True)
(1): Linear(in_features=512, out_features=512, bias=True)
(2): Linear(in_features=512, out_features=2, bias=True)
)
)
)
)
(pos_trans): Linear(in_features=512, out_features=1024, bias=True)
(pos_trans_norm): LayerNorm((1024,), eps=1e-05, elementwise_affine=True)
(reference_points): Linear(in_features=512, out_features=1, bias=True)
)
(caption_head): ModuleList(
(0): LSTMDSACaptioner(
(embed): Embedding(16222, 512)
(logit): Linear(in_features=512, out_features=16222, bias=True)
(dropout): Dropout(p=0.5, inplace=False)
(core): ShowAttendTellCore(
(rnn): LSTM(1536, 512, bias=False, dropout=0.5)
(att_drop): Dropout(p=0.5, inplace=False)
(deformable_att): MSDeformAttnCap(
(sampling_offsets): Linear(in_features=1024, out_features=16, bias=True)
(attention_weights): Linear(in_features=1024, out_features=16, bias=True)
(value_proj): Linear(in_features=512, out_features=512, bias=True)
(output_proj): Linear(in_features=512, out_features=512, bias=True)
)
(ctx2att): Linear(in_features=512, out_features=512, bias=True)
(h2att): Linear(in_features=512, out_features=512, bias=True)
(alpha_net): Linear(in_features=512, out_features=1, bias=True)
)
)
(1): LSTMDSACaptioner(
(embed): Embedding(16222, 512)
(logit): Linear(in_features=512, out_features=16222, bias=True)
(dropout): Dropout(p=0.5, inplace=False)
(core): ShowAttendTellCore(
(rnn): LSTM(1536, 512, bias=False, dropout=0.5)
(att_drop): Dropout(p=0.5, inplace=False)
(deformable_att): MSDeformAttnCap(
(sampling_offsets): Linear(in_features=1024, out_features=16, bias=True)
(attention_weights): Linear(in_features=1024, out_features=16, bias=True)
(value_proj): Linear(in_features=512, out_features=512, bias=True)
(output_proj): Linear(in_features=512, out_features=512, bias=True)
)
(ctx2att): Linear(in_features=512, out_features=512, bias=True)
(h2att): Linear(in_features=512, out_features=512, bias=True)
(alpha_net): Linear(in_features=512, out_features=1, bias=True)
)
)
)
(query_embed): Embedding(100, 1024)
(class_head): ModuleList(
(0): Linear(in_features=512, out_features=1, bias=True)
(1): Linear(in_features=512, out_features=1, bias=True)
)
(class_refine_head): ModuleList(
(0): Linear(in_features=512, out_features=1, bias=True)
(1): Linear(in_features=512, out_features=1, bias=True)
)
(count_head): ModuleList(
(0): Linear(in_features=512, out_features=11, bias=True)
(1): Linear(in_features=512, out_features=11, bias=True)
)
(bbox_head): ModuleList(
(0): MLP(
(layers): ModuleList(
(0): Linear(in_features=512, out_features=512, bias=True)
(1): Linear(in_features=512, out_features=512, bias=True)
(2): Linear(in_features=512, out_features=2, bias=True)
)
)
(1): MLP(
(layers): ModuleList(
(0): Linear(in_features=512, out_features=512, bias=True)
(1): Linear(in_features=512, out_features=512, bias=True)
(2): Linear(in_features=512, out_features=2, bias=True)
)
)
)
(contrastive_projection_event): ModuleList(
(0): Identity()
(1): Identity()
)
(contrastive_projection_text): ModuleList(
(0): Identity()
(1): Identity()
)
)
******************** Strat training ! ******************************************
loss type: dict_keys(['loss_ce', 'loss_bbox', 'loss_giou', 'loss_counter', 'loss_caption', 'contrastive_loss', 'loss_ce_0', 'loss_bbox_0', 'loss_giou_0', 'loss_counter_0', 'loss_caption_0', 'contrastive_loss_0'])
loss weights: dict_values([2, 0, 4, 0.5, 2, 0.0, 2, 0, 4, 0.5, 2, 0.0])
ID seq2-ft(mix)-gt_percent-1.0 iter 1000 (epoch 0),
loss = OrderedDict([('loss_ce', 0.284), ('loss_counter', 0.126), ('loss_bbox', 0.117), ('loss_giou', 0.275), ('loss_self_iou', 0.126), ('cardinality_error', 3.775), ('loss_ce_0', 0.284), ('loss_counter_0', 0.126), ('loss_bbox_0', 0.118), ('loss_giou_0', 0.276), ('loss_self_iou_0', 0.126), ('cardinality_error_0', 3.775), ('loss_caption_0', 3.781), ('loss_caption', 3.778), ('total_loss', 18.585)]),
time/iter = 0.182, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 2000 (epoch 0),
loss = OrderedDict([('loss_ce', 0.287), ('loss_counter', 0.119), ('loss_bbox', 0.087), ('loss_giou', 0.239), ('loss_self_iou', 0.12), ('cardinality_error', 3.705), ('loss_ce_0', 0.289), ('loss_counter_0', 0.118), ('loss_bbox_0', 0.087), ('loss_giou_0', 0.239), ('loss_self_iou_0', 0.121), ('cardinality_error_0', 3.705), ('loss_caption_0', 3.682), ('loss_caption', 3.675), ('total_loss', 17.896)]),
time/iter = 0.180, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 3000 (epoch 0),
loss = OrderedDict([('loss_ce', 0.291), ('loss_counter', 0.122), ('loss_bbox', 0.078), ('loss_giou', 0.227), ('loss_self_iou', 0.098), ('cardinality_error', 3.705), ('loss_ce_0', 0.292), ('loss_counter_0', 0.122), ('loss_bbox_0', 0.078), ('loss_giou_0', 0.228), ('loss_self_iou_0', 0.099), ('cardinality_error_0', 3.705), ('loss_caption_0', 3.668), ('loss_caption', 3.664), ('total_loss', 17.771)]),
time/iter = 0.181, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 4000 (epoch 0),
loss = OrderedDict([('loss_ce', 0.289), ('loss_counter', 0.126), ('loss_bbox', 0.078), ('loss_giou', 0.224), ('loss_self_iou', 0.1), ('cardinality_error', 3.784), ('loss_ce_0', 0.291), ('loss_counter_0', 0.127), ('loss_bbox_0', 0.078), ('loss_giou_0', 0.223), ('loss_self_iou_0', 0.1), ('cardinality_error_0', 3.784), ('loss_caption_0', 3.624), ('loss_caption', 3.629), ('total_loss', 17.579)]),
time/iter = 0.174, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 5000 (epoch 0),
loss = OrderedDict([('loss_ce', 0.285), ('loss_counter', 0.121), ('loss_bbox', 0.08), ('loss_giou', 0.218), ('loss_self_iou', 0.114), ('cardinality_error', 3.674), ('loss_ce_0', 0.287), ('loss_counter_0', 0.121), ('loss_bbox_0', 0.08), ('loss_giou_0', 0.218), ('loss_self_iou_0', 0.115), ('cardinality_error_0', 3.674), ('loss_caption_0', 3.629), ('loss_caption', 3.629), ('total_loss', 17.526)]),
time/iter = 0.178, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 6000 (epoch 0),
loss = OrderedDict([('loss_ce', 0.292), ('loss_counter', 0.13), ('loss_bbox', 0.076), ('loss_giou', 0.22), ('loss_self_iou', 0.098), ('cardinality_error', 3.786), ('loss_ce_0', 0.293), ('loss_counter_0', 0.129), ('loss_bbox_0', 0.076), ('loss_giou_0', 0.22), ('loss_self_iou_0', 0.098), ('cardinality_error_0', 3.786), ('loss_caption_0', 3.625), ('loss_caption', 3.622), ('total_loss', 17.555)]),
time/iter = 0.182, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 7000 (epoch 0),
loss = OrderedDict([('loss_ce', 0.292), ('loss_counter', 0.12), ('loss_bbox', 0.076), ('loss_giou', 0.215), ('loss_self_iou', 0.097), ('cardinality_error', 3.746), ('loss_ce_0', 0.293), ('loss_counter_0', 0.119), ('loss_bbox_0', 0.076), ('loss_giou_0', 0.215), ('loss_self_iou_0', 0.099), ('cardinality_error_0', 3.746), ('loss_caption_0', 3.58), ('loss_caption', 3.576), ('total_loss', 17.319)]),
time/iter = 0.179, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 8000 (epoch 0),
loss = OrderedDict([('loss_ce', 0.288), ('loss_counter', 0.129), ('loss_bbox', 0.078), ('loss_giou', 0.218), ('loss_self_iou', 0.108), ('cardinality_error', 3.754), ('loss_ce_0', 0.288), ('loss_counter_0', 0.128), ('loss_bbox_0', 0.079), ('loss_giou_0', 0.218), ('loss_self_iou_0', 0.11), ('cardinality_error_0', 3.754), ('loss_caption_0', 3.546), ('loss_caption', 3.546), ('total_loss', 17.209)]),
time/iter = 0.184, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 9000 (epoch 0),
loss = OrderedDict([('loss_ce', 0.29), ('loss_counter', 0.12), ('loss_bbox', 0.078), ('loss_giou', 0.219), ('loss_self_iou', 0.1), ('cardinality_error', 3.685), ('loss_ce_0', 0.291), ('loss_counter_0', 0.12), ('loss_bbox_0', 0.078), ('loss_giou_0', 0.219), ('loss_self_iou_0', 0.101), ('cardinality_error_0', 3.685), ('loss_caption_0', 3.544), ('loss_caption', 3.54), ('total_loss', 17.2)]),
time/iter = 0.180, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 10000 (epoch 0),
loss = OrderedDict([('loss_ce', 0.293), ('loss_counter', 0.125), ('loss_bbox', 0.077), ('loss_giou', 0.22), ('loss_self_iou', 0.101), ('cardinality_error', 3.748), ('loss_ce_0', 0.293), ('loss_counter_0', 0.125), ('loss_bbox_0', 0.078), ('loss_giou_0', 0.22), ('loss_self_iou_0', 0.102), ('cardinality_error_0', 3.748), ('loss_caption_0', 3.582), ('loss_caption', 3.577), ('total_loss', 17.376)]),
time/iter = 0.180, bad_vid = 0.000
Validation results of iter 10009:
Bleu_1:0.15656016917085527
Bleu_2:0.08210369852679855
Bleu_3:0.042491746140277446
Bleu_4:0.021149866989626908
METEOR:0.08752782819459405
ROUGE_L:0.1577032846084498
CIDEr:0.2687260839927409
Recall:0.4986985069085389
Precision:0.548450952477792
soda_c:0.045070258467165024
para_Bleu_1:0.36987086578065714
para_Bleu_2:0.1987998709052068
para_Bleu_3:0.11671522868501899
para_Bleu_4:0.07164097958462183
para_METEOR:0.13901753612789455
para_ROUGE_L:0.2826680559963382
para_CIDEr:0.0956891322121665
overall score of iter 10009: 0.3063476479246829
Save model at iter 10009 to /mnt/data/pjlab-3090-sport/wuhao/logs/dibs/howto-anet_anet_ori_pbox(similarity_op_order_v2)_CLIP/similarity_op_order_v2_topf30_iter3_r1_th1_refine_aug(8,0.02)_top3_2stage_ins_cap_topk_mil_coef0_noFocal_seq2-ft(mix)-gt_percent-1.0_2/model-last.pth.
Save Best-model at iter 10009 to checkpoint file.
Save info to info.json
ID seq2-ft(mix)-gt_percent-1.0 iter 11000 (epoch 1),
loss = OrderedDict([('loss_ce', 0.29), ('loss_counter', 0.124), ('loss_bbox', 0.077), ('loss_giou', 0.217), ('loss_self_iou', 0.101), ('cardinality_error', 3.788), ('loss_ce_0', 0.292), ('loss_counter_0', 0.123), ('loss_bbox_0', 0.076), ('loss_giou_0', 0.217), ('loss_self_iou_0', 0.099), ('cardinality_error_0', 3.788), ('loss_caption_0', 3.446), ('loss_caption', 3.443), ('total_loss', 16.802)]),
time/iter = 0.707, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 12000 (epoch 1),
loss = OrderedDict([('loss_ce', 0.29), ('loss_counter', 0.12), ('loss_bbox', 0.076), ('loss_giou', 0.214), ('loss_self_iou', 0.103), ('cardinality_error', 3.694), ('loss_ce_0', 0.291), ('loss_counter_0', 0.12), ('loss_bbox_0', 0.075), ('loss_giou_0', 0.213), ('loss_self_iou_0', 0.103), ('cardinality_error_0', 3.694), ('loss_caption_0', 3.427), ('loss_caption', 3.428), ('total_loss', 16.701)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 13000 (epoch 1),
loss = OrderedDict([('loss_ce', 0.291), ('loss_counter', 0.12), ('loss_bbox', 0.076), ('loss_giou', 0.217), ('loss_self_iou', 0.107), ('cardinality_error', 3.689), ('loss_ce_0', 0.291), ('loss_counter_0', 0.12), ('loss_bbox_0', 0.076), ('loss_giou_0', 0.217), ('loss_self_iou_0', 0.107), ('cardinality_error_0', 3.689), ('loss_caption_0', 3.464), ('loss_caption', 3.461), ('total_loss', 16.871)]),
time/iter = 0.189, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 14000 (epoch 1),
loss = OrderedDict([('loss_ce', 0.292), ('loss_counter', 0.118), ('loss_bbox', 0.073), ('loss_giou', 0.21), ('loss_self_iou', 0.1), ('cardinality_error', 3.663), ('loss_ce_0', 0.292), ('loss_counter_0', 0.118), ('loss_bbox_0', 0.073), ('loss_giou_0', 0.211), ('loss_self_iou_0', 0.101), ('cardinality_error_0', 3.663), ('loss_caption_0', 3.414), ('loss_caption', 3.41), ('total_loss', 16.616)]),
time/iter = 0.189, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 15000 (epoch 1),
loss = OrderedDict([('loss_ce', 0.295), ('loss_counter', 0.127), ('loss_bbox', 0.076), ('loss_giou', 0.214), ('loss_self_iou', 0.103), ('cardinality_error', 3.828), ('loss_ce_0', 0.296), ('loss_counter_0', 0.127), ('loss_bbox_0', 0.076), ('loss_giou_0', 0.215), ('loss_self_iou_0', 0.104), ('cardinality_error_0', 3.828), ('loss_caption_0', 3.453), ('loss_caption', 3.453), ('total_loss', 16.836)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 16000 (epoch 1),
loss = OrderedDict([('loss_ce', 0.296), ('loss_counter', 0.121), ('loss_bbox', 0.073), ('loss_giou', 0.206), ('loss_self_iou', 0.105), ('cardinality_error', 3.687), ('loss_ce_0', 0.297), ('loss_counter_0', 0.12), ('loss_bbox_0', 0.072), ('loss_giou_0', 0.207), ('loss_self_iou_0', 0.104), ('cardinality_error_0', 3.687), ('loss_caption_0', 3.461), ('loss_caption', 3.462), ('total_loss', 16.803)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 17000 (epoch 1),
loss = OrderedDict([('loss_ce', 0.3), ('loss_counter', 0.127), ('loss_bbox', 0.073), ('loss_giou', 0.208), ('loss_self_iou', 0.102), ('cardinality_error', 3.791), ('loss_ce_0', 0.3), ('loss_counter_0', 0.127), ('loss_bbox_0', 0.073), ('loss_giou_0', 0.209), ('loss_self_iou_0', 0.103), ('cardinality_error_0', 3.791), ('loss_caption_0', 3.469), ('loss_caption', 3.465), ('total_loss', 16.864)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 18000 (epoch 1),
loss = OrderedDict([('loss_ce', 0.298), ('loss_counter', 0.119), ('loss_bbox', 0.074), ('loss_giou', 0.205), ('loss_self_iou', 0.107), ('cardinality_error', 3.68), ('loss_ce_0', 0.298), ('loss_counter_0', 0.119), ('loss_bbox_0', 0.074), ('loss_giou_0', 0.206), ('loss_self_iou_0', 0.107), ('cardinality_error_0', 3.68), ('loss_caption_0', 3.478), ('loss_caption', 3.475), ('total_loss', 16.859)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 19000 (epoch 1),
loss = OrderedDict([('loss_ce', 0.305), ('loss_counter', 0.126), ('loss_bbox', 0.073), ('loss_giou', 0.207), ('loss_self_iou', 0.099), ('cardinality_error', 3.752), ('loss_ce_0', 0.304), ('loss_counter_0', 0.126), ('loss_bbox_0', 0.072), ('loss_giou_0', 0.208), ('loss_self_iou_0', 0.099), ('cardinality_error_0', 3.752), ('loss_caption_0', 3.396), ('loss_caption', 3.396), ('total_loss', 16.585)]),
time/iter = 0.194, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 20000 (epoch 1),
loss = OrderedDict([('loss_ce', 0.303), ('loss_counter', 0.128), ('loss_bbox', 0.071), ('loss_giou', 0.208), ('loss_self_iou', 0.101), ('cardinality_error', 3.804), ('loss_ce_0', 0.304), ('loss_counter_0', 0.128), ('loss_bbox_0', 0.071), ('loss_giou_0', 0.208), ('loss_self_iou_0', 0.1), ('cardinality_error_0', 3.804), ('loss_caption_0', 3.42), ('loss_caption', 3.419), ('total_loss', 16.684)]),
time/iter = 0.189, bad_vid = 0.000
Validation results of iter 20018:
Bleu_1:0.15965966113561106
Bleu_2:0.08785069799970043
Bleu_3:0.04739925348589703
Bleu_4:0.02377096308421814
METEOR:0.09062964515721111
ROUGE_L:0.1652647774491388
CIDEr:0.27366191469495676
Recall:0.45131293652113946
Precision:0.5379414954918249
soda_c:0.04303682007432423
para_Bleu_1:0.3640361416830845
para_Bleu_2:0.1986476696673755
para_Bleu_3:0.11814800235116821
para_Bleu_4:0.07336184523852665
para_METEOR:0.13911724177507803
para_ROUGE_L:0.28211794880017504
para_CIDEr:0.08634617454158834
overall score of iter 20018: 0.29882526155519307
Save model at iter 20018 to /mnt/data/pjlab-3090-sport/wuhao/logs/dibs/howto-anet_anet_ori_pbox(similarity_op_order_v2)_CLIP/similarity_op_order_v2_topf30_iter3_r1_th1_refine_aug(8,0.02)_top3_2stage_ins_cap_topk_mil_coef0_noFocal_seq2-ft(mix)-gt_percent-1.0_2/model-last.pth.
Save info to info.json
ID seq2-ft(mix)-gt_percent-1.0 iter 21000 (epoch 2),
loss = OrderedDict([('loss_ce', 0.298), ('loss_counter', 0.122), ('loss_bbox', 0.071), ('loss_giou', 0.202), ('loss_self_iou', 0.101), ('cardinality_error', 3.666), ('loss_ce_0', 0.299), ('loss_counter_0', 0.122), ('loss_bbox_0', 0.071), ('loss_giou_0', 0.202), ('loss_self_iou_0', 0.103), ('cardinality_error_0', 3.666), ('loss_caption_0', 3.344), ('loss_caption', 3.335), ('total_loss', 16.294)]),
time/iter = 0.726, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 22000 (epoch 2),
loss = OrderedDict([('loss_ce', 0.293), ('loss_counter', 0.119), ('loss_bbox', 0.073), ('loss_giou', 0.201), ('loss_self_iou', 0.109), ('cardinality_error', 3.752), ('loss_ce_0', 0.292), ('loss_counter_0', 0.118), ('loss_bbox_0', 0.073), ('loss_giou_0', 0.203), ('loss_self_iou_0', 0.11), ('cardinality_error_0', 3.752), ('loss_caption_0', 3.302), ('loss_caption', 3.304), ('total_loss', 16.116)]),
time/iter = 0.188, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 23000 (epoch 2),
loss = OrderedDict([('loss_ce', 0.299), ('loss_counter', 0.128), ('loss_bbox', 0.077), ('loss_giou', 0.208), ('loss_self_iou', 0.113), ('cardinality_error', 3.803), ('loss_ce_0', 0.299), ('loss_counter_0', 0.128), ('loss_bbox_0', 0.076), ('loss_giou_0', 0.208), ('loss_self_iou_0', 0.112), ('cardinality_error_0', 3.803), ('loss_caption_0', 3.348), ('loss_caption', 3.34), ('total_loss', 16.363)]),
time/iter = 0.190, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 24000 (epoch 2),
loss = OrderedDict([('loss_ce', 0.293), ('loss_counter', 0.122), ('loss_bbox', 0.076), ('loss_giou', 0.207), ('loss_self_iou', 0.093), ('cardinality_error', 3.729), ('loss_ce_0', 0.294), ('loss_counter_0', 0.122), ('loss_bbox_0', 0.076), ('loss_giou_0', 0.207), ('loss_self_iou_0', 0.094), ('cardinality_error_0', 3.729), ('loss_caption_0', 3.354), ('loss_caption', 3.351), ('total_loss', 16.364)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 25000 (epoch 2),
loss = OrderedDict([('loss_ce', 0.294), ('loss_counter', 0.122), ('loss_bbox', 0.078), ('loss_giou', 0.213), ('loss_self_iou', 0.091), ('cardinality_error', 3.734), ('loss_ce_0', 0.295), ('loss_counter_0', 0.122), ('loss_bbox_0', 0.077), ('loss_giou_0', 0.214), ('loss_self_iou_0', 0.09), ('cardinality_error_0', 3.734), ('loss_caption_0', 3.372), ('loss_caption', 3.372), ('total_loss', 16.494)]),
time/iter = 0.192, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 26000 (epoch 2),
loss = OrderedDict([('loss_ce', 0.298), ('loss_counter', 0.125), ('loss_bbox', 0.072), ('loss_giou', 0.203), ('loss_self_iou', 0.096), ('cardinality_error', 3.784), ('loss_ce_0', 0.299), ('loss_counter_0', 0.125), ('loss_bbox_0', 0.073), ('loss_giou_0', 0.204), ('loss_self_iou_0', 0.096), ('cardinality_error_0', 3.784), ('loss_caption_0', 3.334), ('loss_caption', 3.333), ('total_loss', 16.279)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 27000 (epoch 2),
loss = OrderedDict([('loss_ce', 0.289), ('loss_counter', 0.118), ('loss_bbox', 0.076), ('loss_giou', 0.203), ('loss_self_iou', 0.102), ('cardinality_error', 3.64), ('loss_ce_0', 0.291), ('loss_counter_0', 0.119), ('loss_bbox_0', 0.076), ('loss_giou_0', 0.203), ('loss_self_iou_0', 0.101), ('cardinality_error_0', 3.64), ('loss_caption_0', 3.348), ('loss_caption', 3.345), ('total_loss', 16.287)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 28000 (epoch 2),
loss = OrderedDict([('loss_ce', 0.292), ('loss_counter', 0.125), ('loss_bbox', 0.077), ('loss_giou', 0.201), ('loss_self_iou', 0.095), ('cardinality_error', 3.774), ('loss_ce_0', 0.293), ('loss_counter_0', 0.125), ('loss_bbox_0', 0.076), ('loss_giou_0', 0.202), ('loss_self_iou_0', 0.096), ('cardinality_error_0', 3.774), ('loss_caption_0', 3.337), ('loss_caption', 3.333), ('total_loss', 16.249)]),
time/iter = 0.194, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 29000 (epoch 2),
loss = OrderedDict([('loss_ce', 0.298), ('loss_counter', 0.12), ('loss_bbox', 0.075), ('loss_giou', 0.204), ('loss_self_iou', 0.1), ('cardinality_error', 3.755), ('loss_ce_0', 0.299), ('loss_counter_0', 0.12), ('loss_bbox_0', 0.074), ('loss_giou_0', 0.205), ('loss_self_iou_0', 0.101), ('cardinality_error_0', 3.755), ('loss_caption_0', 3.315), ('loss_caption', 3.321), ('total_loss', 16.223)]),
time/iter = 0.192, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 30000 (epoch 2),
loss = OrderedDict([('loss_ce', 0.302), ('loss_counter', 0.119), ('loss_bbox', 0.071), ('loss_giou', 0.195), ('loss_self_iou', 0.103), ('cardinality_error', 3.72), ('loss_ce_0', 0.302), ('loss_counter_0', 0.119), ('loss_bbox_0', 0.072), ('loss_giou_0', 0.196), ('loss_self_iou_0', 0.104), ('cardinality_error_0', 3.72), ('loss_caption_0', 3.347), ('loss_caption', 3.349), ('total_loss', 16.283)]),
time/iter = 0.195, bad_vid = 0.000
Validation results of iter 30027:
Bleu_1:0.15440507165989542
Bleu_2:0.08178273697953425
Bleu_3:0.042600749568780155
Bleu_4:0.02119123483046711
METEOR:0.08563216148714695
ROUGE_L:0.156809182143994
CIDEr:0.25960752079137744
Recall:0.5075951227720545
Precision:0.571834112941489
soda_c:0.048597974030683
para_Bleu_1:0.3985431504573892
para_Bleu_2:0.22415947108296613
para_Bleu_3:0.1341003834690626
para_Bleu_4:0.08312155143550452
para_METEOR:0.1510085678983445
para_ROUGE_L:0.2957598062989384
para_CIDEr:0.12271570278513648
overall score of iter 30027: 0.3568458221189855
Save model at iter 30027 to /mnt/data/pjlab-3090-sport/wuhao/logs/dibs/howto-anet_anet_ori_pbox(similarity_op_order_v2)_CLIP/similarity_op_order_v2_topf30_iter3_r1_th1_refine_aug(8,0.02)_top3_2stage_ins_cap_topk_mil_coef0_noFocal_seq2-ft(mix)-gt_percent-1.0_2/model-last.pth.
Save Best-model at iter 30027 to checkpoint file.
Save info to info.json
ID seq2-ft(mix)-gt_percent-1.0 iter 31000 (epoch 3),
loss = OrderedDict([('loss_ce', 0.296), ('loss_counter', 0.123), ('loss_bbox', 0.073), ('loss_giou', 0.202), ('loss_self_iou', 0.114), ('cardinality_error', 3.772), ('loss_ce_0', 0.296), ('loss_counter_0', 0.123), ('loss_bbox_0', 0.074), ('loss_giou_0', 0.203), ('loss_self_iou_0', 0.115), ('cardinality_error_0', 3.772), ('loss_caption_0', 3.24), ('loss_caption', 3.242), ('total_loss', 15.889)]),
time/iter = 0.725, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 32000 (epoch 3),
loss = OrderedDict([('loss_ce', 0.3), ('loss_counter', 0.117), ('loss_bbox', 0.069), ('loss_giou', 0.193), ('loss_self_iou', 0.093), ('cardinality_error', 3.66), ('loss_ce_0', 0.3), ('loss_counter_0', 0.117), ('loss_bbox_0', 0.07), ('loss_giou_0', 0.195), ('loss_self_iou_0', 0.092), ('cardinality_error_0', 3.66), ('loss_caption_0', 3.251), ('loss_caption', 3.248), ('total_loss', 15.869)]),
time/iter = 0.192, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 33000 (epoch 3),
loss = OrderedDict([('loss_ce', 0.302), ('loss_counter', 0.126), ('loss_bbox', 0.07), ('loss_giou', 0.197), ('loss_self_iou', 0.102), ('cardinality_error', 3.787), ('loss_ce_0', 0.301), ('loss_counter_0', 0.126), ('loss_bbox_0', 0.071), ('loss_giou_0', 0.199), ('loss_self_iou_0', 0.104), ('cardinality_error_0', 3.787), ('loss_caption_0', 3.223), ('loss_caption', 3.225), ('total_loss', 15.81)]),
time/iter = 0.189, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 34000 (epoch 3),
loss = OrderedDict([('loss_ce', 0.297), ('loss_counter', 0.121), ('loss_bbox', 0.076), ('loss_giou', 0.201), ('loss_self_iou', 0.107), ('cardinality_error', 3.719), ('loss_ce_0', 0.296), ('loss_counter_0', 0.121), ('loss_bbox_0', 0.077), ('loss_giou_0', 0.202), ('loss_self_iou_0', 0.108), ('cardinality_error_0', 3.719), ('loss_caption_0', 3.21), ('loss_caption', 3.206), ('total_loss', 15.752)]),
time/iter = 0.189, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 35000 (epoch 3),
loss = OrderedDict([('loss_ce', 0.303), ('loss_counter', 0.122), ('loss_bbox', 0.074), ('loss_giou', 0.201), ('loss_self_iou', 0.1), ('cardinality_error', 3.761), ('loss_ce_0', 0.304), ('loss_counter_0', 0.121), ('loss_bbox_0', 0.073), ('loss_giou_0', 0.202), ('loss_self_iou_0', 0.1), ('cardinality_error_0', 3.761), ('loss_caption_0', 3.261), ('loss_caption', 3.267), ('total_loss', 16.006)]),
time/iter = 0.187, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 36000 (epoch 3),
loss = OrderedDict([('loss_ce', 0.302), ('loss_counter', 0.12), ('loss_bbox', 0.074), ('loss_giou', 0.202), ('loss_self_iou', 0.096), ('cardinality_error', 3.731), ('loss_ce_0', 0.302), ('loss_counter_0', 0.12), ('loss_bbox_0', 0.075), ('loss_giou_0', 0.203), ('loss_self_iou_0', 0.096), ('cardinality_error_0', 3.731), ('loss_caption_0', 3.322), ('loss_caption', 3.322), ('total_loss', 16.237)]),
time/iter = 0.192, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 37000 (epoch 3),
loss = OrderedDict([('loss_ce', 0.306), ('loss_counter', 0.12), ('loss_bbox', 0.069), ('loss_giou', 0.193), ('loss_self_iou', 0.088), ('cardinality_error', 3.747), ('loss_ce_0', 0.306), ('loss_counter_0', 0.12), ('loss_bbox_0', 0.069), ('loss_giou_0', 0.195), ('loss_self_iou_0', 0.089), ('cardinality_error_0', 3.747), ('loss_caption_0', 3.276), ('loss_caption', 3.278), ('total_loss', 16.005)]),
time/iter = 0.194, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 38000 (epoch 3),
loss = OrderedDict([('loss_ce', 0.295), ('loss_counter', 0.122), ('loss_bbox', 0.073), ('loss_giou', 0.198), ('loss_self_iou', 0.096), ('cardinality_error', 3.747), ('loss_ce_0', 0.295), ('loss_counter_0', 0.122), ('loss_bbox_0', 0.074), ('loss_giou_0', 0.199), ('loss_self_iou_0', 0.096), ('cardinality_error_0', 3.747), ('loss_caption_0', 3.26), ('loss_caption', 3.267), ('total_loss', 15.944)]),
time/iter = 0.192, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 39000 (epoch 3),
loss = OrderedDict([('loss_ce', 0.301), ('loss_counter', 0.12), ('loss_bbox', 0.073), ('loss_giou', 0.194), ('loss_self_iou', 0.096), ('cardinality_error', 3.714), ('loss_ce_0', 0.3), ('loss_counter_0', 0.12), ('loss_bbox_0', 0.074), ('loss_giou_0', 0.196), ('loss_self_iou_0', 0.098), ('cardinality_error_0', 3.714), ('loss_caption_0', 3.29), ('loss_caption', 3.284), ('total_loss', 16.029)]),
time/iter = 0.192, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 40000 (epoch 3),
loss = OrderedDict([('loss_ce', 0.302), ('loss_counter', 0.124), ('loss_bbox', 0.068), ('loss_giou', 0.187), ('loss_self_iou', 0.098), ('cardinality_error', 3.742), ('loss_ce_0', 0.302), ('loss_counter_0', 0.124), ('loss_bbox_0', 0.069), ('loss_giou_0', 0.189), ('loss_self_iou_0', 0.099), ('cardinality_error_0', 3.742), ('loss_caption_0', 3.255), ('loss_caption', 3.258), ('total_loss', 15.861)]),
time/iter = 0.191, bad_vid = 0.000
Validation results of iter 40036:
Bleu_1:0.16003947012491918
Bleu_2:0.08640386650819816
Bleu_3:0.045769192920880976
Bleu_4:0.023139762266241797
METEOR:0.08893476927946467
ROUGE_L:0.16285119298911696
CIDEr:0.27850058398714506
Recall:0.4974410652224822
Precision:0.571762083926507
soda_c:0.04898353247531122
para_Bleu_1:0.4116267700746525
para_Bleu_2:0.23315066082372427
para_Bleu_3:0.139785630195007
para_Bleu_4:0.08689414164874545
para_METEOR:0.15321412716959742
para_ROUGE_L:0.2993749803089721
para_CIDEr:0.12755194391496638
overall score of iter 40036: 0.3676602127333093
Save model at iter 40036 to /mnt/data/pjlab-3090-sport/wuhao/logs/dibs/howto-anet_anet_ori_pbox(similarity_op_order_v2)_CLIP/similarity_op_order_v2_topf30_iter3_r1_th1_refine_aug(8,0.02)_top3_2stage_ins_cap_topk_mil_coef0_noFocal_seq2-ft(mix)-gt_percent-1.0_2/model-last.pth.
Save Best-model at iter 40036 to checkpoint file.
Save info to info.json
ID seq2-ft(mix)-gt_percent-1.0 iter 41000 (epoch 4),
loss = OrderedDict([('loss_ce', 0.304), ('loss_counter', 0.122), ('loss_bbox', 0.071), ('loss_giou', 0.196), ('loss_self_iou', 0.094), ('cardinality_error', 3.73), ('loss_ce_0', 0.303), ('loss_counter_0', 0.121), ('loss_bbox_0', 0.071), ('loss_giou_0', 0.197), ('loss_self_iou_0', 0.095), ('cardinality_error_0', 3.73), ('loss_caption_0', 3.159), ('loss_caption', 3.162), ('total_loss', 15.549)]),
time/iter = 0.733, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 42000 (epoch 4),
loss = OrderedDict([('loss_ce', 0.297), ('loss_counter', 0.117), ('loss_bbox', 0.072), ('loss_giou', 0.188), ('loss_self_iou', 0.097), ('cardinality_error', 3.698), ('loss_ce_0', 0.298), ('loss_counter_0', 0.116), ('loss_bbox_0', 0.071), ('loss_giou_0', 0.189), ('loss_self_iou_0', 0.096), ('cardinality_error_0', 3.698), ('loss_caption_0', 3.191), ('loss_caption', 3.187), ('total_loss', 15.571)]),
time/iter = 0.194, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 43000 (epoch 4),
loss = OrderedDict([('loss_ce', 0.306), ('loss_counter', 0.12), ('loss_bbox', 0.07), ('loss_giou', 0.198), ('loss_self_iou', 0.089), ('cardinality_error', 3.785), ('loss_ce_0', 0.306), ('loss_counter_0', 0.119), ('loss_bbox_0', 0.069), ('loss_giou_0', 0.2), ('loss_self_iou_0', 0.087), ('cardinality_error_0', 3.785), ('loss_caption_0', 3.247), ('loss_caption', 3.249), ('total_loss', 15.93)]),
time/iter = 0.195, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 44000 (epoch 4),
loss = OrderedDict([('loss_ce', 0.301), ('loss_counter', 0.12), ('loss_bbox', 0.072), ('loss_giou', 0.194), ('loss_self_iou', 0.104), ('cardinality_error', 3.727), ('loss_ce_0', 0.302), ('loss_counter_0', 0.12), ('loss_bbox_0', 0.072), ('loss_giou_0', 0.195), ('loss_self_iou_0', 0.102), ('cardinality_error_0', 3.727), ('loss_caption_0', 3.228), ('loss_caption', 3.227), ('total_loss', 15.794)]),
time/iter = 0.192, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 45000 (epoch 4),
loss = OrderedDict([('loss_ce', 0.303), ('loss_counter', 0.12), ('loss_bbox', 0.07), ('loss_giou', 0.194), ('loss_self_iou', 0.094), ('cardinality_error', 3.684), ('loss_ce_0', 0.304), ('loss_counter_0', 0.12), ('loss_bbox_0', 0.07), ('loss_giou_0', 0.196), ('loss_self_iou_0', 0.094), ('cardinality_error_0', 3.684), ('loss_caption_0', 3.138), ('loss_caption', 3.143), ('total_loss', 15.458)]),
time/iter = 0.190, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 46000 (epoch 4),
loss = OrderedDict([('loss_ce', 0.302), ('loss_counter', 0.123), ('loss_bbox', 0.071), ('loss_giou', 0.194), ('loss_self_iou', 0.107), ('cardinality_error', 3.8), ('loss_ce_0', 0.301), ('loss_counter_0', 0.122), ('loss_bbox_0', 0.071), ('loss_giou_0', 0.196), ('loss_self_iou_0', 0.107), ('cardinality_error_0', 3.8), ('loss_caption_0', 3.198), ('loss_caption', 3.202), ('total_loss', 15.69)]),
time/iter = 0.192, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 47000 (epoch 4),
loss = OrderedDict([('loss_ce', 0.302), ('loss_counter', 0.124), ('loss_bbox', 0.071), ('loss_giou', 0.193), ('loss_self_iou', 0.1), ('cardinality_error', 3.724), ('loss_ce_0', 0.302), ('loss_counter_0', 0.123), ('loss_bbox_0', 0.072), ('loss_giou_0', 0.194), ('loss_self_iou_0', 0.101), ('cardinality_error_0', 3.724), ('loss_caption_0', 3.166), ('loss_caption', 3.167), ('total_loss', 15.544)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 48000 (epoch 4),
loss = OrderedDict([('loss_ce', 0.302), ('loss_counter', 0.126), ('loss_bbox', 0.074), ('loss_giou', 0.194), ('loss_self_iou', 0.1), ('cardinality_error', 3.779), ('loss_ce_0', 0.303), ('loss_counter_0', 0.126), ('loss_bbox_0', 0.073), ('loss_giou_0', 0.195), ('loss_self_iou_0', 0.099), ('cardinality_error_0', 3.779), ('loss_caption_0', 3.197), ('loss_caption', 3.204), ('total_loss', 15.693)]),
time/iter = 0.194, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 49000 (epoch 4),
loss = OrderedDict([('loss_ce', 0.3), ('loss_counter', 0.117), ('loss_bbox', 0.072), ('loss_giou', 0.186), ('loss_self_iou', 0.103), ('cardinality_error', 3.67), ('loss_ce_0', 0.299), ('loss_counter_0', 0.117), ('loss_bbox_0', 0.073), ('loss_giou_0', 0.189), ('loss_self_iou_0', 0.103), ('cardinality_error_0', 3.67), ('loss_caption_0', 3.197), ('loss_caption', 3.193), ('total_loss', 15.597)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 50000 (epoch 4),
loss = OrderedDict([('loss_ce', 0.303), ('loss_counter', 0.122), ('loss_bbox', 0.071), ('loss_giou', 0.191), ('loss_self_iou', 0.1), ('cardinality_error', 3.769), ('loss_ce_0', 0.303), ('loss_counter_0', 0.121), ('loss_bbox_0', 0.07), ('loss_giou_0', 0.192), ('loss_self_iou_0', 0.098), ('cardinality_error_0', 3.769), ('loss_caption_0', 3.195), ('loss_caption', 3.196), ('total_loss', 15.646)]),
time/iter = 0.193, bad_vid = 0.000
Validation results of iter 50045:
Bleu_1:0.1612752203314224
Bleu_2:0.08712092952271142
Bleu_3:0.04643407984417907
Bleu_4:0.024237450149938583
METEOR:0.0888552980469009
ROUGE_L:0.16165678007821221
CIDEr:0.28844655875134945
Recall:0.5079771255793173
Precision:0.5707494407158785
soda_c:0.05143467092505771
para_Bleu_1:0.425828341023263
para_Bleu_2:0.2431293051387748
para_Bleu_3:0.14662751878582
para_Bleu_4:0.09131956416083617
para_METEOR:0.15868276543147294
para_ROUGE_L:0.30762031965083425
para_CIDEr:0.1438790695271004
overall score of iter 50045: 0.39388139911940956
Save model at iter 50045 to /mnt/data/pjlab-3090-sport/wuhao/logs/dibs/howto-anet_anet_ori_pbox(similarity_op_order_v2)_CLIP/similarity_op_order_v2_topf30_iter3_r1_th1_refine_aug(8,0.02)_top3_2stage_ins_cap_topk_mil_coef0_noFocal_seq2-ft(mix)-gt_percent-1.0_2/model-last.pth.
Save Best-model at iter 50045 to checkpoint file.
Save info to info.json
ID seq2-ft(mix)-gt_percent-1.0 iter 51000 (epoch 5),
loss = OrderedDict([('loss_ce', 0.304), ('loss_counter', 0.119), ('loss_bbox', 0.072), ('loss_giou', 0.19), ('loss_self_iou', 0.1), ('cardinality_error', 3.708), ('loss_ce_0', 0.304), ('loss_counter_0', 0.119), ('loss_bbox_0', 0.07), ('loss_giou_0', 0.19), ('loss_self_iou_0', 0.099), ('cardinality_error_0', 3.708), ('loss_caption_0', 3.123), ('loss_caption', 3.122), ('total_loss', 15.345)]),
time/iter = 0.739, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 52000 (epoch 5),
loss = OrderedDict([('loss_ce', 0.302), ('loss_counter', 0.122), ('loss_bbox', 0.07), ('loss_giou', 0.195), ('loss_self_iou', 0.091), ('cardinality_error', 3.787), ('loss_ce_0', 0.302), ('loss_counter_0', 0.121), ('loss_bbox_0', 0.07), ('loss_giou_0', 0.198), ('loss_self_iou_0', 0.092), ('cardinality_error_0', 3.787), ('loss_caption_0', 3.08), ('loss_caption', 3.08), ('total_loss', 15.224)]),
time/iter = 0.193, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 53000 (epoch 5),
loss = OrderedDict([('loss_ce', 0.303), ('loss_counter', 0.12), ('loss_bbox', 0.07), ('loss_giou', 0.192), ('loss_self_iou', 0.101), ('cardinality_error', 3.688), ('loss_ce_0', 0.302), ('loss_counter_0', 0.12), ('loss_bbox_0', 0.071), ('loss_giou_0', 0.194), ('loss_self_iou_0', 0.102), ('cardinality_error_0', 3.688), ('loss_caption_0', 3.121), ('loss_caption', 3.125), ('total_loss', 15.366)]),
time/iter = 0.196, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 54000 (epoch 5),
loss = OrderedDict([('loss_ce', 0.304), ('loss_counter', 0.12), ('loss_bbox', 0.069), ('loss_giou', 0.184), ('loss_self_iou', 0.096), ('cardinality_error', 3.66), ('loss_ce_0', 0.303), ('loss_counter_0', 0.12), ('loss_bbox_0', 0.07), ('loss_giou_0', 0.187), ('loss_self_iou_0', 0.096), ('cardinality_error_0', 3.66), ('loss_caption_0', 3.151), ('loss_caption', 3.158), ('total_loss', 15.44)]),
time/iter = 0.193, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 55000 (epoch 5),
loss = OrderedDict([('loss_ce', 0.314), ('loss_counter', 0.123), ('loss_bbox', 0.069), ('loss_giou', 0.186), ('loss_self_iou', 0.102), ('cardinality_error', 3.759), ('loss_ce_0', 0.314), ('loss_counter_0', 0.124), ('loss_bbox_0', 0.069), ('loss_giou_0', 0.188), ('loss_self_iou_0', 0.103), ('cardinality_error_0', 3.759), ('loss_caption_0', 3.137), ('loss_caption', 3.138), ('total_loss', 15.427)]),
time/iter = 0.193, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 56000 (epoch 5),
loss = OrderedDict([('loss_ce', 0.304), ('loss_counter', 0.12), ('loss_bbox', 0.069), ('loss_giou', 0.186), ('loss_self_iou', 0.102), ('cardinality_error', 3.7), ('loss_ce_0', 0.303), ('loss_counter_0', 0.119), ('loss_bbox_0', 0.07), ('loss_giou_0', 0.189), ('loss_self_iou_0', 0.102), ('cardinality_error_0', 3.7), ('loss_caption_0', 3.128), ('loss_caption', 3.132), ('total_loss', 15.353)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 57000 (epoch 5),
loss = OrderedDict([('loss_ce', 0.308), ('loss_counter', 0.125), ('loss_bbox', 0.069), ('loss_giou', 0.192), ('loss_self_iou', 0.094), ('cardinality_error', 3.833), ('loss_ce_0', 0.308), ('loss_counter_0', 0.125), ('loss_bbox_0', 0.069), ('loss_giou_0', 0.194), ('loss_self_iou_0', 0.093), ('cardinality_error_0', 3.833), ('loss_caption_0', 3.157), ('loss_caption', 3.154), ('total_loss', 15.516)]),
time/iter = 0.192, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 58000 (epoch 5),
loss = OrderedDict([('loss_ce', 0.3), ('loss_counter', 0.116), ('loss_bbox', 0.072), ('loss_giou', 0.192), ('loss_self_iou', 0.099), ('cardinality_error', 3.724), ('loss_ce_0', 0.3), ('loss_counter_0', 0.116), ('loss_bbox_0', 0.073), ('loss_giou_0', 0.192), ('loss_self_iou_0', 0.1), ('cardinality_error_0', 3.724), ('loss_caption_0', 3.092), ('loss_caption', 3.088), ('total_loss', 15.209)]),
time/iter = 0.189, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 59000 (epoch 5),
loss = OrderedDict([('loss_ce', 0.305), ('loss_counter', 0.126), ('loss_bbox', 0.07), ('loss_giou', 0.187), ('loss_self_iou', 0.092), ('cardinality_error', 3.806), ('loss_ce_0', 0.304), ('loss_counter_0', 0.126), ('loss_bbox_0', 0.07), ('loss_giou_0', 0.19), ('loss_self_iou_0', 0.092), ('cardinality_error_0', 3.806), ('loss_caption_0', 3.204), ('loss_caption', 3.204), ('total_loss', 15.668)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 60000 (epoch 5),
loss = OrderedDict([('loss_ce', 0.298), ('loss_counter', 0.119), ('loss_bbox', 0.073), ('loss_giou', 0.197), ('loss_self_iou', 0.102), ('cardinality_error', 3.73), ('loss_ce_0', 0.298), ('loss_counter_0', 0.118), ('loss_bbox_0', 0.074), ('loss_giou_0', 0.198), ('loss_self_iou_0', 0.101), ('cardinality_error_0', 3.73), ('loss_caption_0', 3.185), ('loss_caption', 3.179), ('total_loss', 15.62)]),
time/iter = 0.192, bad_vid = 0.000
Validation results of iter 60054:
Bleu_1:0.16203040821313286
Bleu_2:0.087418866671477
Bleu_3:0.04641401855891123
Bleu_4:0.023872355329811287
METEOR:0.08736154709181514
ROUGE_L:0.16095171754962678
CIDEr:0.3019460931650574
Recall:0.5237442505746305
Precision:0.5691986983933232
soda_c:0.05366939846142926
para_Bleu_1:0.4285515683378188
para_Bleu_2:0.24896313523930838
para_Bleu_3:0.15083849533584295
para_Bleu_4:0.09425440122753082
para_METEOR:0.15418242275887206
para_ROUGE_L:0.3037081433191389
para_CIDEr:0.16822639157343386
overall score of iter 60054: 0.41666321555983676
Save model at iter 60054 to /mnt/data/pjlab-3090-sport/wuhao/logs/dibs/howto-anet_anet_ori_pbox(similarity_op_order_v2)_CLIP/similarity_op_order_v2_topf30_iter3_r1_th1_refine_aug(8,0.02)_top3_2stage_ins_cap_topk_mil_coef0_noFocal_seq2-ft(mix)-gt_percent-1.0_2/model-last.pth.
Save Best-model at iter 60054 to checkpoint file.
Save info to info.json
ID seq2-ft(mix)-gt_percent-1.0 iter 61000 (epoch 6),
loss = OrderedDict([('loss_ce', 0.302), ('loss_counter', 0.117), ('loss_bbox', 0.068), ('loss_giou', 0.183), ('loss_self_iou', 0.099), ('cardinality_error', 3.687), ('loss_ce_0', 0.303), ('loss_counter_0', 0.117), ('loss_bbox_0', 0.067), ('loss_giou_0', 0.185), ('loss_self_iou_0', 0.1), ('cardinality_error_0', 3.687), ('loss_caption_0', 3.025), ('loss_caption', 3.031), ('total_loss', 14.914)]),
time/iter = 0.715, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 62000 (epoch 6),
loss = OrderedDict([('loss_ce', 0.305), ('loss_counter', 0.125), ('loss_bbox', 0.068), ('loss_giou', 0.192), ('loss_self_iou', 0.088), ('cardinality_error', 3.809), ('loss_ce_0', 0.304), ('loss_counter_0', 0.125), ('loss_bbox_0', 0.069), ('loss_giou_0', 0.194), ('loss_self_iou_0', 0.089), ('cardinality_error_0', 3.809), ('loss_caption_0', 3.067), ('loss_caption', 3.064), ('total_loss', 15.147)]),
time/iter = 0.192, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 63000 (epoch 6),
loss = OrderedDict([('loss_ce', 0.301), ('loss_counter', 0.113), ('loss_bbox', 0.072), ('loss_giou', 0.189), ('loss_self_iou', 0.102), ('cardinality_error', 3.636), ('loss_ce_0', 0.301), ('loss_counter_0', 0.113), ('loss_bbox_0', 0.073), ('loss_giou_0', 0.193), ('loss_self_iou_0', 0.104), ('cardinality_error_0', 3.636), ('loss_caption_0', 3.09), ('loss_caption', 3.083), ('total_loss', 15.188)]),
time/iter = 0.192, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 64000 (epoch 6),
loss = OrderedDict([('loss_ce', 0.308), ('loss_counter', 0.12), ('loss_bbox', 0.067), ('loss_giou', 0.185), ('loss_self_iou', 0.105), ('cardinality_error', 3.738), ('loss_ce_0', 0.309), ('loss_counter_0', 0.12), ('loss_bbox_0', 0.067), ('loss_giou_0', 0.186), ('loss_self_iou_0', 0.104), ('cardinality_error_0', 3.738), ('loss_caption_0', 3.09), ('loss_caption', 3.088), ('total_loss', 15.193)]),
time/iter = 0.194, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 65000 (epoch 6),
loss = OrderedDict([('loss_ce', 0.302), ('loss_counter', 0.123), ('loss_bbox', 0.069), ('loss_giou', 0.191), ('loss_self_iou', 0.094), ('cardinality_error', 3.735), ('loss_ce_0', 0.304), ('loss_counter_0', 0.123), ('loss_bbox_0', 0.069), ('loss_giou_0', 0.191), ('loss_self_iou_0', 0.094), ('cardinality_error_0', 3.735), ('loss_caption_0', 3.087), ('loss_caption', 3.083), ('total_loss', 15.203)]),
time/iter = 0.190, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 66000 (epoch 6),
loss = OrderedDict([('loss_ce', 0.307), ('loss_counter', 0.121), ('loss_bbox', 0.069), ('loss_giou', 0.188), ('loss_self_iou', 0.095), ('cardinality_error', 3.753), ('loss_ce_0', 0.307), ('loss_counter_0', 0.121), ('loss_bbox_0', 0.07), ('loss_giou_0', 0.19), ('loss_self_iou_0', 0.095), ('cardinality_error_0', 3.753), ('loss_caption_0', 3.093), ('loss_caption', 3.093), ('total_loss', 15.235)]),
time/iter = 0.192, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 67000 (epoch 6),
loss = OrderedDict([('loss_ce', 0.299), ('loss_counter', 0.123), ('loss_bbox', 0.071), ('loss_giou', 0.189), ('loss_self_iou', 0.099), ('cardinality_error', 3.781), ('loss_ce_0', 0.299), ('loss_counter_0', 0.123), ('loss_bbox_0', 0.072), ('loss_giou_0', 0.192), ('loss_self_iou_0', 0.098), ('cardinality_error_0', 3.781), ('loss_caption_0', 3.104), ('loss_caption', 3.095), ('total_loss', 15.24)]),
time/iter = 0.194, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 68000 (epoch 6),
loss = OrderedDict([('loss_ce', 0.3), ('loss_counter', 0.118), ('loss_bbox', 0.073), ('loss_giou', 0.186), ('loss_self_iou', 0.102), ('cardinality_error', 3.702), ('loss_ce_0', 0.3), ('loss_counter_0', 0.118), ('loss_bbox_0', 0.073), ('loss_giou_0', 0.187), ('loss_self_iou_0', 0.104), ('cardinality_error_0', 3.702), ('loss_caption_0', 3.092), ('loss_caption', 3.087), ('total_loss', 15.171)]),
time/iter = 0.193, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 69000 (epoch 6),
loss = OrderedDict([('loss_ce', 0.304), ('loss_counter', 0.116), ('loss_bbox', 0.068), ('loss_giou', 0.184), ('loss_self_iou', 0.087), ('cardinality_error', 3.705), ('loss_ce_0', 0.303), ('loss_counter_0', 0.116), ('loss_bbox_0', 0.069), ('loss_giou_0', 0.187), ('loss_self_iou_0', 0.088), ('cardinality_error_0', 3.705), ('loss_caption_0', 3.087), ('loss_caption', 3.084), ('total_loss', 15.154)]),
time/iter = 0.192, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 70000 (epoch 6),
loss = OrderedDict([('loss_ce', 0.308), ('loss_counter', 0.119), ('loss_bbox', 0.07), ('loss_giou', 0.188), ('loss_self_iou', 0.104), ('cardinality_error', 3.763), ('loss_ce_0', 0.309), ('loss_counter_0', 0.12), ('loss_bbox_0', 0.069), ('loss_giou_0', 0.19), ('loss_self_iou_0', 0.104), ('cardinality_error_0', 3.763), ('loss_caption_0', 3.137), ('loss_caption', 3.142), ('total_loss', 15.421)]),
time/iter = 0.201, bad_vid = 0.000
Validation results of iter 70063:
Bleu_1:0.17095715677415013
Bleu_2:0.0951967897773989
Bleu_3:0.05145074727592996
Bleu_4:0.026686223548170303
METEOR:0.09033289555302068
ROUGE_L:0.16939818741017104
CIDEr:0.33299543538258497
Recall:0.5001550726802355
Precision:0.5629321740898863
soda_c:0.05378783144134501
para_Bleu_1:0.44719474980697405
para_Bleu_2:0.2615784516531111
para_Bleu_3:0.15956746990786394
para_Bleu_4:0.09983770060804388
para_METEOR:0.15549284849496958
para_ROUGE_L:0.30852597622578265
para_CIDEr:0.18758102150887232
overall score of iter 70063: 0.4429115706118858
Save model at iter 70063 to /mnt/data/pjlab-3090-sport/wuhao/logs/dibs/howto-anet_anet_ori_pbox(similarity_op_order_v2)_CLIP/similarity_op_order_v2_topf30_iter3_r1_th1_refine_aug(8,0.02)_top3_2stage_ins_cap_topk_mil_coef0_noFocal_seq2-ft(mix)-gt_percent-1.0_2/model-last.pth.
Save Best-model at iter 70063 to checkpoint file.
Save info to info.json
ID seq2-ft(mix)-gt_percent-1.0 iter 71000 (epoch 7),
loss = OrderedDict([('loss_ce', 0.304), ('loss_counter', 0.115), ('loss_bbox', 0.067), ('loss_giou', 0.187), ('loss_self_iou', 0.091), ('cardinality_error', 3.724), ('loss_ce_0', 0.304), ('loss_counter_0', 0.115), ('loss_bbox_0', 0.068), ('loss_giou_0', 0.189), ('loss_self_iou_0', 0.092), ('cardinality_error_0', 3.724), ('loss_caption_0', 2.994), ('loss_caption', 2.994), ('total_loss', 14.812)]),
time/iter = 0.691, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 72000 (epoch 7),
loss = OrderedDict([('loss_ce', 0.297), ('loss_counter', 0.118), ('loss_bbox', 0.07), ('loss_giou', 0.187), ('loss_self_iou', 0.099), ('cardinality_error', 3.665), ('loss_ce_0', 0.296), ('loss_counter_0', 0.118), ('loss_bbox_0', 0.072), ('loss_giou_0', 0.19), ('loss_self_iou_0', 0.099), ('cardinality_error_0', 3.665), ('loss_caption_0', 2.995), ('loss_caption', 3.0), ('total_loss', 14.803)]),
time/iter = 0.193, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 73000 (epoch 7),
loss = OrderedDict([('loss_ce', 0.301), ('loss_counter', 0.122), ('loss_bbox', 0.067), ('loss_giou', 0.183), ('loss_self_iou', 0.099), ('cardinality_error', 3.762), ('loss_ce_0', 0.302), ('loss_counter_0', 0.122), ('loss_bbox_0', 0.067), ('loss_giou_0', 0.184), ('loss_self_iou_0', 0.099), ('cardinality_error_0', 3.762), ('loss_caption_0', 3.03), ('loss_caption', 3.034), ('total_loss', 14.924)]),
time/iter = 0.189, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 74000 (epoch 7),
loss = OrderedDict([('loss_ce', 0.303), ('loss_counter', 0.12), ('loss_bbox', 0.067), ('loss_giou', 0.181), ('loss_self_iou', 0.093), ('cardinality_error', 3.722), ('loss_ce_0', 0.304), ('loss_counter_0', 0.12), ('loss_bbox_0', 0.068), ('loss_giou_0', 0.183), ('loss_self_iou_0', 0.095), ('cardinality_error_0', 3.722), ('loss_caption_0', 3.061), ('loss_caption', 3.062), ('total_loss', 15.037)]),
time/iter = 0.192, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 75000 (epoch 7),
loss = OrderedDict([('loss_ce', 0.3), ('loss_counter', 0.124), ('loss_bbox', 0.069), ('loss_giou', 0.188), ('loss_self_iou', 0.097), ('cardinality_error', 3.835), ('loss_ce_0', 0.302), ('loss_counter_0', 0.124), ('loss_bbox_0', 0.069), ('loss_giou_0', 0.19), ('loss_self_iou_0', 0.097), ('cardinality_error_0', 3.835), ('loss_caption_0', 3.102), ('loss_caption', 3.108), ('total_loss', 15.261)]),
time/iter = 0.195, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 76000 (epoch 7),
loss = OrderedDict([('loss_ce', 0.304), ('loss_counter', 0.118), ('loss_bbox', 0.069), ('loss_giou', 0.19), ('loss_self_iou', 0.096), ('cardinality_error', 3.787), ('loss_ce_0', 0.305), ('loss_counter_0', 0.118), ('loss_bbox_0', 0.069), ('loss_giou_0', 0.192), ('loss_self_iou_0', 0.096), ('cardinality_error_0', 3.787), ('loss_caption_0', 3.055), ('loss_caption', 3.056), ('total_loss', 15.081)]),
time/iter = 0.192, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 77000 (epoch 7),
loss = OrderedDict([('loss_ce', 0.3), ('loss_counter', 0.122), ('loss_bbox', 0.07), ('loss_giou', 0.191), ('loss_self_iou', 0.101), ('cardinality_error', 3.753), ('loss_ce_0', 0.3), ('loss_counter_0', 0.122), ('loss_bbox_0', 0.071), ('loss_giou_0', 0.192), ('loss_self_iou_0', 0.102), ('cardinality_error_0', 3.753), ('loss_caption_0', 3.064), ('loss_caption', 3.063), ('total_loss', 15.105)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 78000 (epoch 7),
loss = OrderedDict([('loss_ce', 0.303), ('loss_counter', 0.118), ('loss_bbox', 0.069), ('loss_giou', 0.192), ('loss_self_iou', 0.094), ('cardinality_error', 3.812), ('loss_ce_0', 0.302), ('loss_counter_0', 0.118), ('loss_bbox_0', 0.071), ('loss_giou_0', 0.194), ('loss_self_iou_0', 0.095), ('cardinality_error_0', 3.812), ('loss_caption_0', 3.075), ('loss_caption', 3.081), ('total_loss', 15.186)]),
time/iter = 0.192, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 79000 (epoch 7),
loss = OrderedDict([('loss_ce', 0.303), ('loss_counter', 0.119), ('loss_bbox', 0.068), ('loss_giou', 0.184), ('loss_self_iou', 0.099), ('cardinality_error', 3.712), ('loss_ce_0', 0.304), ('loss_counter_0', 0.119), ('loss_bbox_0', 0.068), ('loss_giou_0', 0.187), ('loss_self_iou_0', 0.1), ('cardinality_error_0', 3.712), ('loss_caption_0', 3.004), ('loss_caption', 3.004), ('total_loss', 14.833)]),
time/iter = 0.189, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 80000 (epoch 7),
loss = OrderedDict([('loss_ce', 0.297), ('loss_counter', 0.117), ('loss_bbox', 0.068), ('loss_giou', 0.184), ('loss_self_iou', 0.099), ('cardinality_error', 3.639), ('loss_ce_0', 0.298), ('loss_counter_0', 0.117), ('loss_bbox_0', 0.069), ('loss_giou_0', 0.185), ('loss_self_iou_0', 0.099), ('cardinality_error_0', 3.639), ('loss_caption_0', 3.011), ('loss_caption', 3.021), ('total_loss', 14.846)]),
time/iter = 0.189, bad_vid = 0.000
Validation results of iter 80072:
Bleu_1:0.16525493799366836
Bleu_2:0.09017429361474327
Bleu_3:0.04843073565357156
Bleu_4:0.025752141227780294
METEOR:0.09042668571725655
ROUGE_L:0.1657835735936403
CIDEr:0.30766696683798356
Recall:0.5070758476264831
Precision:0.5698723815334497
soda_c:0.05193286444599829
para_Bleu_1:0.4299765573510605
para_Bleu_2:0.24998607326423264
para_Bleu_3:0.15168978606887273
para_Bleu_4:0.09540463753102806
para_METEOR:0.15913054274631774
para_ROUGE_L:0.30821511076520103
para_CIDEr:0.14655297481419807
overall score of iter 80072: 0.4010881550915439
Save model at iter 80072 to /mnt/data/pjlab-3090-sport/wuhao/logs/dibs/howto-anet_anet_ori_pbox(similarity_op_order_v2)_CLIP/similarity_op_order_v2_topf30_iter3_r1_th1_refine_aug(8,0.02)_top3_2stage_ins_cap_topk_mil_coef0_noFocal_seq2-ft(mix)-gt_percent-1.0_2/model-last.pth.
Save info to info.json
ID seq2-ft(mix)-gt_percent-1.0 iter 81000 (epoch 8),
loss = OrderedDict([('loss_ce', 0.3), ('loss_counter', 0.116), ('loss_bbox', 0.064), ('loss_giou', 0.177), ('loss_self_iou', 0.098), ('cardinality_error', 3.664), ('loss_ce_0', 0.3), ('loss_counter_0', 0.116), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.178), ('loss_self_iou_0', 0.097), ('cardinality_error_0', 3.664), ('loss_caption_0', 2.972), ('loss_caption', 2.974), ('total_loss', 14.63)]),
time/iter = 0.723, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 82000 (epoch 8),
loss = OrderedDict([('loss_ce', 0.301), ('loss_counter', 0.113), ('loss_bbox', 0.067), ('loss_giou', 0.179), ('loss_self_iou', 0.098), ('cardinality_error', 3.692), ('loss_ce_0', 0.301), ('loss_counter_0', 0.113), ('loss_bbox_0', 0.068), ('loss_giou_0', 0.181), ('loss_self_iou_0', 0.099), ('cardinality_error_0', 3.692), ('loss_caption_0', 2.914), ('loss_caption', 2.912), ('total_loss', 14.413)]),
time/iter = 0.190, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 83000 (epoch 8),
loss = OrderedDict([('loss_ce', 0.297), ('loss_counter', 0.117), ('loss_bbox', 0.067), ('loss_giou', 0.188), ('loss_self_iou', 0.097), ('cardinality_error', 3.764), ('loss_ce_0', 0.298), ('loss_counter_0', 0.117), ('loss_bbox_0', 0.068), ('loss_giou_0', 0.19), ('loss_self_iou_0', 0.099), ('cardinality_error_0', 3.764), ('loss_caption_0', 2.939), ('loss_caption', 2.933), ('total_loss', 14.562)]),
time/iter = 0.188, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 84000 (epoch 8),
loss = OrderedDict([('loss_ce', 0.299), ('loss_counter', 0.119), ('loss_bbox', 0.066), ('loss_giou', 0.18), ('loss_self_iou', 0.086), ('cardinality_error', 3.724), ('loss_ce_0', 0.3), ('loss_counter_0', 0.119), ('loss_bbox_0', 0.066), ('loss_giou_0', 0.181), ('loss_self_iou_0', 0.086), ('cardinality_error_0', 3.724), ('loss_caption_0', 2.964), ('loss_caption', 2.963), ('total_loss', 14.614)]),
time/iter = 0.192, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 85000 (epoch 8),
loss = OrderedDict([('loss_ce', 0.301), ('loss_counter', 0.114), ('loss_bbox', 0.066), ('loss_giou', 0.187), ('loss_self_iou', 0.094), ('cardinality_error', 3.73), ('loss_ce_0', 0.301), ('loss_counter_0', 0.114), ('loss_bbox_0', 0.066), ('loss_giou_0', 0.189), ('loss_self_iou_0', 0.095), ('cardinality_error_0', 3.73), ('loss_caption_0', 2.942), ('loss_caption', 2.945), ('total_loss', 14.596)]),
time/iter = 0.190, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 86000 (epoch 8),
loss = OrderedDict([('loss_ce', 0.297), ('loss_counter', 0.118), ('loss_bbox', 0.067), ('loss_giou', 0.184), ('loss_self_iou', 0.096), ('cardinality_error', 3.764), ('loss_ce_0', 0.298), ('loss_counter_0', 0.118), ('loss_bbox_0', 0.068), ('loss_giou_0', 0.187), ('loss_self_iou_0', 0.097), ('cardinality_error_0', 3.764), ('loss_caption_0', 2.989), ('loss_caption', 2.988), ('total_loss', 14.745)]),
time/iter = 0.193, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 87000 (epoch 8),
loss = OrderedDict([('loss_ce', 0.295), ('loss_counter', 0.119), ('loss_bbox', 0.067), ('loss_giou', 0.178), ('loss_self_iou', 0.096), ('cardinality_error', 3.692), ('loss_ce_0', 0.298), ('loss_counter_0', 0.119), ('loss_bbox_0', 0.068), ('loss_giou_0', 0.182), ('loss_self_iou_0', 0.096), ('cardinality_error_0', 3.692), ('loss_caption_0', 2.93), ('loss_caption', 2.931), ('total_loss', 14.465)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 88000 (epoch 8),
loss = OrderedDict([('loss_ce', 0.299), ('loss_counter', 0.117), ('loss_bbox', 0.068), ('loss_giou', 0.181), ('loss_self_iou', 0.102), ('cardinality_error', 3.74), ('loss_ce_0', 0.298), ('loss_counter_0', 0.117), ('loss_bbox_0', 0.07), ('loss_giou_0', 0.184), ('loss_self_iou_0', 0.105), ('cardinality_error_0', 3.74), ('loss_caption_0', 2.945), ('loss_caption', 2.939), ('total_loss', 14.538)]),
time/iter = 0.189, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 89000 (epoch 8),
loss = OrderedDict([('loss_ce', 0.302), ('loss_counter', 0.124), ('loss_bbox', 0.069), ('loss_giou', 0.186), ('loss_self_iou', 0.096), ('cardinality_error', 3.911), ('loss_ce_0', 0.303), ('loss_counter_0', 0.124), ('loss_bbox_0', 0.069), ('loss_giou_0', 0.188), ('loss_self_iou_0', 0.096), ('cardinality_error_0', 3.911), ('loss_caption_0', 2.981), ('loss_caption', 2.985), ('total_loss', 14.762)]),
time/iter = 0.190, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 90000 (epoch 8),
loss = OrderedDict([('loss_ce', 0.298), ('loss_counter', 0.113), ('loss_bbox', 0.066), ('loss_giou', 0.174), ('loss_self_iou', 0.099), ('cardinality_error', 3.667), ('loss_ce_0', 0.3), ('loss_counter_0', 0.112), ('loss_bbox_0', 0.067), ('loss_giou_0', 0.177), ('loss_self_iou_0', 0.099), ('cardinality_error_0', 3.667), ('loss_caption_0', 2.946), ('loss_caption', 2.945), ('total_loss', 14.493)]),
time/iter = 0.191, bad_vid = 0.000
Validation results of iter 90081:
Bleu_1:0.1659435247550983
Bleu_2:0.09010888064116455
Bleu_3:0.04740925434645997
Bleu_4:0.023810200153797586
METEOR:0.0893691583245007
ROUGE_L:0.16481267120708817
CIDEr:0.3096929324572276
Recall:0.5271698247293078
Precision:0.5766981899532185
soda_c:0.05637593299631936
para_Bleu_1:0.4507795558374508
para_Bleu_2:0.2668765313566654
para_Bleu_3:0.16324000259413463
para_Bleu_4:0.10292908422008885
para_METEOR:0.163503434468027
para_ROUGE_L:0.3141109355407807
para_CIDEr:0.1830754815850521
overall score of iter 90081: 0.44950800027316795
Save model at iter 90081 to /mnt/data/pjlab-3090-sport/wuhao/logs/dibs/howto-anet_anet_ori_pbox(similarity_op_order_v2)_CLIP/similarity_op_order_v2_topf30_iter3_r1_th1_refine_aug(8,0.02)_top3_2stage_ins_cap_topk_mil_coef0_noFocal_seq2-ft(mix)-gt_percent-1.0_2/model-last.pth.
Save Best-model at iter 90081 to checkpoint file.
Save info to info.json
ID seq2-ft(mix)-gt_percent-1.0 iter 91000 (epoch 9),
loss = OrderedDict([('loss_ce', 0.296), ('loss_counter', 0.121), ('loss_bbox', 0.066), ('loss_giou', 0.179), ('loss_self_iou', 0.097), ('cardinality_error', 3.807), ('loss_ce_0', 0.298), ('loss_counter_0', 0.12), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.182), ('loss_self_iou_0', 0.097), ('cardinality_error_0', 3.807), ('loss_caption_0', 2.916), ('loss_caption', 2.914), ('total_loss', 14.411)]),
time/iter = 0.724, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 92000 (epoch 9),
loss = OrderedDict([('loss_ce', 0.298), ('loss_counter', 0.121), ('loss_bbox', 0.067), ('loss_giou', 0.179), ('loss_self_iou', 0.093), ('cardinality_error', 3.784), ('loss_ce_0', 0.298), ('loss_counter_0', 0.121), ('loss_bbox_0', 0.068), ('loss_giou_0', 0.182), ('loss_self_iou_0', 0.094), ('cardinality_error_0', 3.784), ('loss_caption_0', 2.916), ('loss_caption', 2.915), ('total_loss', 14.422)]),
time/iter = 0.192, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 93000 (epoch 9),
loss = OrderedDict([('loss_ce', 0.298), ('loss_counter', 0.117), ('loss_bbox', 0.065), ('loss_giou', 0.18), ('loss_self_iou', 0.091), ('cardinality_error', 3.806), ('loss_ce_0', 0.3), ('loss_counter_0', 0.117), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.183), ('loss_self_iou_0', 0.091), ('cardinality_error_0', 3.806), ('loss_caption_0', 2.9), ('loss_caption', 2.905), ('total_loss', 14.377)]),
time/iter = 0.189, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 94000 (epoch 9),
loss = OrderedDict([('loss_ce', 0.293), ('loss_counter', 0.109), ('loss_bbox', 0.068), ('loss_giou', 0.174), ('loss_self_iou', 0.105), ('cardinality_error', 3.616), ('loss_ce_0', 0.293), ('loss_counter_0', 0.109), ('loss_bbox_0', 0.069), ('loss_giou_0', 0.178), ('loss_self_iou_0', 0.106), ('cardinality_error_0', 3.616), ('loss_caption_0', 2.912), ('loss_caption', 2.914), ('total_loss', 14.339)]),
time/iter = 0.187, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 95000 (epoch 9),
loss = OrderedDict([('loss_ce', 0.295), ('loss_counter', 0.12), ('loss_bbox', 0.066), ('loss_giou', 0.185), ('loss_self_iou', 0.093), ('cardinality_error', 3.805), ('loss_ce_0', 0.296), ('loss_counter_0', 0.12), ('loss_bbox_0', 0.068), ('loss_giou_0', 0.187), ('loss_self_iou_0', 0.094), ('cardinality_error_0', 3.805), ('loss_caption_0', 2.938), ('loss_caption', 2.941), ('total_loss', 14.546)]),
time/iter = 0.188, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 96000 (epoch 9),
loss = OrderedDict([('loss_ce', 0.292), ('loss_counter', 0.114), ('loss_bbox', 0.069), ('loss_giou', 0.177), ('loss_self_iou', 0.103), ('cardinality_error', 3.684), ('loss_ce_0', 0.293), ('loss_counter_0', 0.114), ('loss_bbox_0', 0.07), ('loss_giou_0', 0.181), ('loss_self_iou_0', 0.105), ('cardinality_error_0', 3.684), ('loss_caption_0', 2.928), ('loss_caption', 2.931), ('total_loss', 14.434)]),
time/iter = 0.193, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 97000 (epoch 9),
loss = OrderedDict([('loss_ce', 0.297), ('loss_counter', 0.111), ('loss_bbox', 0.066), ('loss_giou', 0.184), ('loss_self_iou', 0.095), ('cardinality_error', 3.693), ('loss_ce_0', 0.298), ('loss_counter_0', 0.111), ('loss_bbox_0', 0.068), ('loss_giou_0', 0.187), ('loss_self_iou_0', 0.097), ('cardinality_error_0', 3.693), ('loss_caption_0', 2.902), ('loss_caption', 2.903), ('total_loss', 14.392)]),
time/iter = 0.190, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 98000 (epoch 9),
loss = OrderedDict([('loss_ce', 0.296), ('loss_counter', 0.115), ('loss_bbox', 0.068), ('loss_giou', 0.181), ('loss_self_iou', 0.089), ('cardinality_error', 3.738), ('loss_ce_0', 0.298), ('loss_counter_0', 0.115), ('loss_bbox_0', 0.068), ('loss_giou_0', 0.184), ('loss_self_iou_0', 0.09), ('cardinality_error_0', 3.738), ('loss_caption_0', 2.896), ('loss_caption', 2.902), ('total_loss', 14.361)]),
time/iter = 0.188, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 99000 (epoch 9),
loss = OrderedDict([('loss_ce', 0.295), ('loss_counter', 0.115), ('loss_bbox', 0.064), ('loss_giou', 0.174), ('loss_self_iou', 0.095), ('cardinality_error', 3.702), ('loss_ce_0', 0.296), ('loss_counter_0', 0.115), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.177), ('loss_self_iou_0', 0.096), ('cardinality_error_0', 3.702), ('loss_caption_0', 2.956), ('loss_caption', 2.956), ('total_loss', 14.525)]),
time/iter = 0.195, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 100000 (epoch 9),
loss = OrderedDict([('loss_ce', 0.296), ('loss_counter', 0.114), ('loss_bbox', 0.066), ('loss_giou', 0.177), ('loss_self_iou', 0.092), ('cardinality_error', 3.751), ('loss_ce_0', 0.298), ('loss_counter_0', 0.113), ('loss_bbox_0', 0.066), ('loss_giou_0', 0.179), ('loss_self_iou_0', 0.093), ('cardinality_error_0', 3.751), ('loss_caption_0', 2.932), ('loss_caption', 2.932), ('total_loss', 14.453)]),
time/iter = 0.191, bad_vid = 0.000
Validation results of iter 100090:
Bleu_1:0.16664911544364056
Bleu_2:0.09023295213839283
Bleu_3:0.04763940550902772
Bleu_4:0.02409205514859969
METEOR:0.0878588871148787
ROUGE_L:0.16401896184386325
CIDEr:0.31947446694949533
Recall:0.5282742157284517
Precision:0.5750796556165633
soda_c:0.05745241491068406
para_Bleu_1:0.46204429574393835
para_Bleu_2:0.2749900961045832
para_Bleu_3:0.1683879565471281
para_Bleu_4:0.10624339593597942
para_METEOR:0.16245439213508253
para_ROUGE_L:0.3162965936511474
para_CIDEr:0.20803178964320856
overall score of iter 100090: 0.4767295777142705
Save model at iter 100090 to /mnt/data/pjlab-3090-sport/wuhao/logs/dibs/howto-anet_anet_ori_pbox(similarity_op_order_v2)_CLIP/similarity_op_order_v2_topf30_iter3_r1_th1_refine_aug(8,0.02)_top3_2stage_ins_cap_topk_mil_coef0_noFocal_seq2-ft(mix)-gt_percent-1.0_2/model-last.pth.
Save Best-model at iter 100090 to checkpoint file.
Save info to info.json
ID seq2-ft(mix)-gt_percent-1.0 iter 101000 (epoch 10),
loss = OrderedDict([('loss_ce', 0.29), ('loss_counter', 0.111), ('loss_bbox', 0.065), ('loss_giou', 0.173), ('loss_self_iou', 0.093), ('cardinality_error', 3.699), ('loss_ce_0', 0.292), ('loss_counter_0', 0.111), ('loss_bbox_0', 0.066), ('loss_giou_0', 0.176), ('loss_self_iou_0', 0.095), ('cardinality_error_0', 3.699), ('loss_caption_0', 2.849), ('loss_caption', 2.847), ('total_loss', 14.064)]),
time/iter = 0.713, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 102000 (epoch 10),
loss = OrderedDict([('loss_ce', 0.292), ('loss_counter', 0.116), ('loss_bbox', 0.065), ('loss_giou', 0.174), ('loss_self_iou', 0.093), ('cardinality_error', 3.695), ('loss_ce_0', 0.293), ('loss_counter_0', 0.117), ('loss_bbox_0', 0.066), ('loss_giou_0', 0.177), ('loss_self_iou_0', 0.094), ('cardinality_error_0', 3.695), ('loss_caption_0', 2.85), ('loss_caption', 2.848), ('total_loss', 14.087)]),
time/iter = 0.189, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 103000 (epoch 10),
loss = OrderedDict([('loss_ce', 0.293), ('loss_counter', 0.115), ('loss_bbox', 0.066), ('loss_giou', 0.173), ('loss_self_iou', 0.093), ('cardinality_error', 3.724), ('loss_ce_0', 0.293), ('loss_counter_0', 0.116), ('loss_bbox_0', 0.067), ('loss_giou_0', 0.178), ('loss_self_iou_0', 0.094), ('cardinality_error_0', 3.724), ('loss_caption_0', 2.846), ('loss_caption', 2.854), ('total_loss', 14.092)]),
time/iter = 0.190, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 104000 (epoch 10),
loss = OrderedDict([('loss_ce', 0.289), ('loss_counter', 0.113), ('loss_bbox', 0.064), ('loss_giou', 0.178), ('loss_self_iou', 0.097), ('cardinality_error', 3.736), ('loss_ce_0', 0.29), ('loss_counter_0', 0.112), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.181), ('loss_self_iou_0', 0.097), ('cardinality_error_0', 3.736), ('loss_caption_0', 2.916), ('loss_caption', 2.913), ('total_loss', 14.362)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 105000 (epoch 10),
loss = OrderedDict([('loss_ce', 0.288), ('loss_counter', 0.117), ('loss_bbox', 0.067), ('loss_giou', 0.18), ('loss_self_iou', 0.091), ('cardinality_error', 3.736), ('loss_ce_0', 0.29), ('loss_counter_0', 0.116), ('loss_bbox_0', 0.068), ('loss_giou_0', 0.183), ('loss_self_iou_0', 0.092), ('cardinality_error_0', 3.736), ('loss_caption_0', 2.907), ('loss_caption', 2.902), ('total_loss', 14.342)]),
time/iter = 0.189, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 106000 (epoch 10),
loss = OrderedDict([('loss_ce', 0.292), ('loss_counter', 0.113), ('loss_bbox', 0.068), ('loss_giou', 0.184), ('loss_self_iou', 0.11), ('cardinality_error', 3.775), ('loss_ce_0', 0.293), ('loss_counter_0', 0.112), ('loss_bbox_0', 0.069), ('loss_giou_0', 0.187), ('loss_self_iou_0', 0.11), ('cardinality_error_0', 3.775), ('loss_caption_0', 2.876), ('loss_caption', 2.875), ('total_loss', 14.264)]),
time/iter = 0.189, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 107000 (epoch 10),
loss = OrderedDict([('loss_ce', 0.291), ('loss_counter', 0.114), ('loss_bbox', 0.069), ('loss_giou', 0.178), ('loss_self_iou', 0.099), ('cardinality_error', 3.743), ('loss_ce_0', 0.291), ('loss_counter_0', 0.114), ('loss_bbox_0', 0.07), ('loss_giou_0', 0.183), ('loss_self_iou_0', 0.101), ('cardinality_error_0', 3.743), ('loss_caption_0', 2.91), ('loss_caption', 2.909), ('total_loss', 14.358)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 108000 (epoch 10),
loss = OrderedDict([('loss_ce', 0.295), ('loss_counter', 0.118), ('loss_bbox', 0.066), ('loss_giou', 0.177), ('loss_self_iou', 0.1), ('cardinality_error', 3.81), ('loss_ce_0', 0.296), ('loss_counter_0', 0.117), ('loss_bbox_0', 0.067), ('loss_giou_0', 0.181), ('loss_self_iou_0', 0.1), ('cardinality_error_0', 3.81), ('loss_caption_0', 2.928), ('loss_caption', 2.93), ('total_loss', 14.446)]),
time/iter = 0.194, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 109000 (epoch 10),
loss = OrderedDict([('loss_ce', 0.294), ('loss_counter', 0.118), ('loss_bbox', 0.063), ('loss_giou', 0.178), ('loss_self_iou', 0.091), ('cardinality_error', 3.78), ('loss_ce_0', 0.296), ('loss_counter_0', 0.117), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.182), ('loss_self_iou_0', 0.092), ('cardinality_error_0', 3.78), ('loss_caption_0', 2.916), ('loss_caption', 2.912), ('total_loss', 14.396)]),
time/iter = 0.193, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 110000 (epoch 10),
loss = OrderedDict([('loss_ce', 0.297), ('loss_counter', 0.113), ('loss_bbox', 0.064), ('loss_giou', 0.178), ('loss_self_iou', 0.087), ('cardinality_error', 3.72), ('loss_ce_0', 0.297), ('loss_counter_0', 0.113), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.184), ('loss_self_iou_0', 0.088), ('cardinality_error_0', 3.72), ('loss_caption_0', 2.948), ('loss_caption', 2.948), ('total_loss', 14.539)]),
time/iter = 0.196, bad_vid = 0.000
Validation results of iter 110099:
Bleu_1:0.1671778590456048
Bleu_2:0.09077014613023152
Bleu_3:0.0476684747303012
Bleu_4:0.02445564298599047
METEOR:0.08933235383587503
ROUGE_L:0.1654660162888944
CIDEr:0.31886265111118334
Recall:0.5314017615268335
Precision:0.5831469052945512
soda_c:0.05853263249839839
para_Bleu_1:0.46544090189732323
para_Bleu_2:0.2789325258737778
para_Bleu_3:0.17172911957785325
para_Bleu_4:0.10903514181091935
para_METEOR:0.16550159188298816
para_ROUGE_L:0.3181118223429575
para_CIDEr:0.2056618808195008
overall score of iter 110099: 0.4801986145134083
Save model at iter 110099 to /mnt/data/pjlab-3090-sport/wuhao/logs/dibs/howto-anet_anet_ori_pbox(similarity_op_order_v2)_CLIP/similarity_op_order_v2_topf30_iter3_r1_th1_refine_aug(8,0.02)_top3_2stage_ins_cap_topk_mil_coef0_noFocal_seq2-ft(mix)-gt_percent-1.0_2/model-last.pth.
Save Best-model at iter 110099 to checkpoint file.
Save info to info.json
ID seq2-ft(mix)-gt_percent-1.0 iter 111000 (epoch 11),
loss = OrderedDict([('loss_ce', 0.286), ('loss_counter', 0.114), ('loss_bbox', 0.066), ('loss_giou', 0.173), ('loss_self_iou', 0.095), ('cardinality_error', 3.718), ('loss_ce_0', 0.287), ('loss_counter_0', 0.113), ('loss_bbox_0', 0.068), ('loss_giou_0', 0.179), ('loss_self_iou_0', 0.097), ('cardinality_error_0', 3.718), ('loss_caption_0', 2.867), ('loss_caption', 2.869), ('total_loss', 14.14)]),
time/iter = 0.727, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 112000 (epoch 11),
loss = OrderedDict([('loss_ce', 0.287), ('loss_counter', 0.111), ('loss_bbox', 0.064), ('loss_giou', 0.169), ('loss_self_iou', 0.098), ('cardinality_error', 3.725), ('loss_ce_0', 0.289), ('loss_counter_0', 0.111), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.176), ('loss_self_iou_0', 0.099), ('cardinality_error_0', 3.725), ('loss_caption_0', 2.844), ('loss_caption', 2.842), ('total_loss', 14.015)]),
time/iter = 0.189, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 113000 (epoch 11),
loss = OrderedDict([('loss_ce', 0.284), ('loss_counter', 0.111), ('loss_bbox', 0.064), ('loss_giou', 0.172), ('loss_self_iou', 0.097), ('cardinality_error', 3.734), ('loss_ce_0', 0.286), ('loss_counter_0', 0.111), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.176), ('loss_self_iou_0', 0.098), ('cardinality_error_0', 3.734), ('loss_caption_0', 2.837), ('loss_caption', 2.834), ('total_loss', 13.981)]),
time/iter = 0.188, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 114000 (epoch 11),
loss = OrderedDict([('loss_ce', 0.283), ('loss_counter', 0.112), ('loss_bbox', 0.064), ('loss_giou', 0.174), ('loss_self_iou', 0.096), ('cardinality_error', 3.739), ('loss_ce_0', 0.285), ('loss_counter_0', 0.111), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.18), ('loss_self_iou_0', 0.096), ('cardinality_error_0', 3.739), ('loss_caption_0', 2.855), ('loss_caption', 2.857), ('total_loss', 14.084)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 115000 (epoch 11),
loss = OrderedDict([('loss_ce', 0.284), ('loss_counter', 0.111), ('loss_bbox', 0.064), ('loss_giou', 0.175), ('loss_self_iou', 0.092), ('cardinality_error', 3.74), ('loss_ce_0', 0.284), ('loss_counter_0', 0.111), ('loss_bbox_0', 0.066), ('loss_giou_0', 0.18), ('loss_self_iou_0', 0.093), ('cardinality_error_0', 3.74), ('loss_caption_0', 2.823), ('loss_caption', 2.824), ('total_loss', 13.959)]),
time/iter = 0.190, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 116000 (epoch 11),
loss = OrderedDict([('loss_ce', 0.286), ('loss_counter', 0.113), ('loss_bbox', 0.065), ('loss_giou', 0.177), ('loss_self_iou', 0.088), ('cardinality_error', 3.753), ('loss_ce_0', 0.288), ('loss_counter_0', 0.113), ('loss_bbox_0', 0.066), ('loss_giou_0', 0.181), ('loss_self_iou_0', 0.088), ('cardinality_error_0', 3.753), ('loss_caption_0', 2.846), ('loss_caption', 2.843), ('total_loss', 14.073)]),
time/iter = 0.192, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 117000 (epoch 11),
loss = OrderedDict([('loss_ce', 0.285), ('loss_counter', 0.113), ('loss_bbox', 0.064), ('loss_giou', 0.174), ('loss_self_iou', 0.096), ('cardinality_error', 3.755), ('loss_ce_0', 0.287), ('loss_counter_0', 0.113), ('loss_bbox_0', 0.064), ('loss_giou_0', 0.179), ('loss_self_iou_0', 0.097), ('cardinality_error_0', 3.755), ('loss_caption_0', 2.804), ('loss_caption', 2.81), ('total_loss', 13.896)]),
time/iter = 0.189, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 118000 (epoch 11),
loss = OrderedDict([('loss_ce', 0.284), ('loss_counter', 0.109), ('loss_bbox', 0.066), ('loss_giou', 0.175), ('loss_self_iou', 0.093), ('cardinality_error', 3.715), ('loss_ce_0', 0.285), ('loss_counter_0', 0.108), ('loss_bbox_0', 0.068), ('loss_giou_0', 0.181), ('loss_self_iou_0', 0.094), ('cardinality_error_0', 3.715), ('loss_caption_0', 2.863), ('loss_caption', 2.866), ('total_loss', 14.129)]),
time/iter = 0.193, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 119000 (epoch 11),
loss = OrderedDict([('loss_ce', 0.286), ('loss_counter', 0.114), ('loss_bbox', 0.064), ('loss_giou', 0.176), ('loss_self_iou', 0.098), ('cardinality_error', 3.735), ('loss_ce_0', 0.287), ('loss_counter_0', 0.114), ('loss_bbox_0', 0.066), ('loss_giou_0', 0.181), ('loss_self_iou_0', 0.098), ('cardinality_error_0', 3.735), ('loss_caption_0', 2.844), ('loss_caption', 2.843), ('total_loss', 14.061)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 120000 (epoch 11),
loss = OrderedDict([('loss_ce', 0.284), ('loss_counter', 0.113), ('loss_bbox', 0.065), ('loss_giou', 0.175), ('loss_self_iou', 0.101), ('cardinality_error', 3.755), ('loss_ce_0', 0.285), ('loss_counter_0', 0.113), ('loss_bbox_0', 0.068), ('loss_giou_0', 0.181), ('loss_self_iou_0', 0.102), ('cardinality_error_0', 3.755), ('loss_caption_0', 2.868), ('loss_caption', 2.878), ('total_loss', 14.168)]),
time/iter = 0.190, bad_vid = 0.000
Validation results of iter 120108:
Bleu_1:0.16560019346009094
Bleu_2:0.08934946581658681
Bleu_3:0.04692472826903507
Bleu_4:0.023331060597699706
METEOR:0.08861943572471001
ROUGE_L:0.16392659155605854
CIDEr:0.31177527957257306
Recall:0.5248955646301546
Precision:0.5713061826316813
soda_c:0.056694173808073595
para_Bleu_1:0.45551540477127933
para_Bleu_2:0.2725270289009415
para_Bleu_3:0.16731081427102573
para_Bleu_4:0.10555679460767188
para_METEOR:0.1665724805603667
para_ROUGE_L:0.31619749898051375
para_CIDEr:0.19719071969736374
overall score of iter 120108: 0.4693199948654023
Save model at iter 120108 to /mnt/data/pjlab-3090-sport/wuhao/logs/dibs/howto-anet_anet_ori_pbox(similarity_op_order_v2)_CLIP/similarity_op_order_v2_topf30_iter3_r1_th1_refine_aug(8,0.02)_top3_2stage_ins_cap_topk_mil_coef0_noFocal_seq2-ft(mix)-gt_percent-1.0_2/model-last.pth.
Save info to info.json
ID seq2-ft(mix)-gt_percent-1.0 iter 121000 (epoch 12),
loss = OrderedDict([('loss_ce', 0.283), ('loss_counter', 0.108), ('loss_bbox', 0.063), ('loss_giou', 0.166), ('loss_self_iou', 0.095), ('cardinality_error', 3.691), ('loss_ce_0', 0.284), ('loss_counter_0', 0.108), ('loss_bbox_0', 0.066), ('loss_giou_0', 0.174), ('loss_self_iou_0', 0.096), ('cardinality_error_0', 3.691), ('loss_caption_0', 2.809), ('loss_caption', 2.808), ('total_loss', 13.835)]),
time/iter = 0.727, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 122000 (epoch 12),
loss = OrderedDict([('loss_ce', 0.28), ('loss_counter', 0.109), ('loss_bbox', 0.064), ('loss_giou', 0.17), ('loss_self_iou', 0.093), ('cardinality_error', 3.706), ('loss_ce_0', 0.281), ('loss_counter_0', 0.108), ('loss_bbox_0', 0.066), ('loss_giou_0', 0.177), ('loss_self_iou_0', 0.093), ('cardinality_error_0', 3.706), ('loss_caption_0', 2.811), ('loss_caption', 2.814), ('total_loss', 13.867)]),
time/iter = 0.190, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 123000 (epoch 12),
loss = OrderedDict([('loss_ce', 0.28), ('loss_counter', 0.109), ('loss_bbox', 0.066), ('loss_giou', 0.172), ('loss_self_iou', 0.097), ('cardinality_error', 3.691), ('loss_ce_0', 0.281), ('loss_counter_0', 0.11), ('loss_bbox_0', 0.067), ('loss_giou_0', 0.179), ('loss_self_iou_0', 0.097), ('cardinality_error_0', 3.691), ('loss_caption_0', 2.789), ('loss_caption', 2.797), ('total_loss', 13.808)]),
time/iter = 0.190, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 124000 (epoch 12),
loss = OrderedDict([('loss_ce', 0.282), ('loss_counter', 0.112), ('loss_bbox', 0.063), ('loss_giou', 0.17), ('loss_self_iou', 0.092), ('cardinality_error', 3.76), ('loss_ce_0', 0.281), ('loss_counter_0', 0.112), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.176), ('loss_self_iou_0', 0.093), ('cardinality_error_0', 3.76), ('loss_caption_0', 2.839), ('loss_caption', 2.842), ('total_loss', 13.984)]),
time/iter = 0.194, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 125000 (epoch 12),
loss = OrderedDict([('loss_ce', 0.281), ('loss_counter', 0.112), ('loss_bbox', 0.064), ('loss_giou', 0.174), ('loss_self_iou', 0.097), ('cardinality_error', 3.763), ('loss_ce_0', 0.282), ('loss_counter_0', 0.112), ('loss_bbox_0', 0.066), ('loss_giou_0', 0.179), ('loss_self_iou_0', 0.097), ('cardinality_error_0', 3.763), ('loss_caption_0', 2.81), ('loss_caption', 2.815), ('total_loss', 13.898)]),
time/iter = 0.189, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 126000 (epoch 12),
loss = OrderedDict([('loss_ce', 0.282), ('loss_counter', 0.112), ('loss_bbox', 0.064), ('loss_giou', 0.177), ('loss_self_iou', 0.095), ('cardinality_error', 3.717), ('loss_ce_0', 0.283), ('loss_counter_0', 0.112), ('loss_bbox_0', 0.066), ('loss_giou_0', 0.183), ('loss_self_iou_0', 0.098), ('cardinality_error_0', 3.717), ('loss_caption_0', 2.789), ('loss_caption', 2.787), ('total_loss', 13.835)]),
time/iter = 0.189, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 127000 (epoch 12),
loss = OrderedDict([('loss_ce', 0.277), ('loss_counter', 0.112), ('loss_bbox', 0.064), ('loss_giou', 0.172), ('loss_self_iou', 0.097), ('cardinality_error', 3.764), ('loss_ce_0', 0.277), ('loss_counter_0', 0.112), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.178), ('loss_self_iou_0', 0.096), ('cardinality_error_0', 3.764), ('loss_caption_0', 2.867), ('loss_caption', 2.871), ('total_loss', 14.097)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 128000 (epoch 12),
loss = OrderedDict([('loss_ce', 0.281), ('loss_counter', 0.113), ('loss_bbox', 0.063), ('loss_giou', 0.173), ('loss_self_iou', 0.092), ('cardinality_error', 3.793), ('loss_ce_0', 0.283), ('loss_counter_0', 0.112), ('loss_bbox_0', 0.064), ('loss_giou_0', 0.179), ('loss_self_iou_0', 0.092), ('cardinality_error_0', 3.793), ('loss_caption_0', 2.868), ('loss_caption', 2.863), ('total_loss', 14.111)]),
time/iter = 0.190, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 129000 (epoch 12),
loss = OrderedDict([('loss_ce', 0.279), ('loss_counter', 0.106), ('loss_bbox', 0.066), ('loss_giou', 0.175), ('loss_self_iou', 0.1), ('cardinality_error', 3.686), ('loss_ce_0', 0.283), ('loss_counter_0', 0.105), ('loss_bbox_0', 0.068), ('loss_giou_0', 0.181), ('loss_self_iou_0', 0.101), ('cardinality_error_0', 3.686), ('loss_caption_0', 2.812), ('loss_caption', 2.813), ('total_loss', 13.903)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 130000 (epoch 12),
loss = OrderedDict([('loss_ce', 0.283), ('loss_counter', 0.111), ('loss_bbox', 0.065), ('loss_giou', 0.174), ('loss_self_iou', 0.097), ('cardinality_error', 3.772), ('loss_ce_0', 0.286), ('loss_counter_0', 0.111), ('loss_bbox_0', 0.066), ('loss_giou_0', 0.179), ('loss_self_iou_0', 0.098), ('cardinality_error_0', 3.772), ('loss_caption_0', 2.86), ('loss_caption', 2.861), ('total_loss', 14.105)]),
time/iter = 0.190, bad_vid = 0.000
Validation results of iter 130117:
Bleu_1:0.16778675341331784
Bleu_2:0.09082555766488616
Bleu_3:0.047445681271689716
Bleu_4:0.02375280793420285
METEOR:0.08883520478698428
ROUGE_L:0.16531435721130755
CIDEr:0.31778343902267087
Recall:0.5273619026669621
Precision:0.5698181479221706
soda_c:0.05753856798988932
para_Bleu_1:0.4610381779339771
para_Bleu_2:0.2761144617772928
para_Bleu_3:0.16915034097081671
para_Bleu_4:0.10654029953240575
para_METEOR:0.16638305166981465
para_ROUGE_L:0.31710573495570465
para_CIDEr:0.19601570682645908
overall score of iter 130117: 0.46893905802867947
Save model at iter 130117 to /mnt/data/pjlab-3090-sport/wuhao/logs/dibs/howto-anet_anet_ori_pbox(similarity_op_order_v2)_CLIP/similarity_op_order_v2_topf30_iter3_r1_th1_refine_aug(8,0.02)_top3_2stage_ins_cap_topk_mil_coef0_noFocal_seq2-ft(mix)-gt_percent-1.0_2/model-last.pth.
Save info to info.json
ID seq2-ft(mix)-gt_percent-1.0 iter 131000 (epoch 13),
loss = OrderedDict([('loss_ce', 0.277), ('loss_counter', 0.107), ('loss_bbox', 0.062), ('loss_giou', 0.17), ('loss_self_iou', 0.092), ('cardinality_error', 3.75), ('loss_ce_0', 0.279), ('loss_counter_0', 0.107), ('loss_bbox_0', 0.064), ('loss_giou_0', 0.178), ('loss_self_iou_0', 0.093), ('cardinality_error_0', 3.75), ('loss_caption_0', 2.817), ('loss_caption', 2.826), ('total_loss', 13.897)]),
time/iter = 0.734, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 132000 (epoch 13),
loss = OrderedDict([('loss_ce', 0.271), ('loss_counter', 0.109), ('loss_bbox', 0.065), ('loss_giou', 0.174), ('loss_self_iou', 0.089), ('cardinality_error', 3.814), ('loss_ce_0', 0.274), ('loss_counter_0', 0.109), ('loss_bbox_0', 0.066), ('loss_giou_0', 0.181), ('loss_self_iou_0', 0.09), ('cardinality_error_0', 3.814), ('loss_caption_0', 2.778), ('loss_caption', 2.776), ('total_loss', 13.726)]),
time/iter = 0.189, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 133000 (epoch 13),
loss = OrderedDict([('loss_ce', 0.277), ('loss_counter', 0.113), ('loss_bbox', 0.064), ('loss_giou', 0.172), ('loss_self_iou', 0.095), ('cardinality_error', 3.773), ('loss_ce_0', 0.277), ('loss_counter_0', 0.112), ('loss_bbox_0', 0.066), ('loss_giou_0', 0.179), ('loss_self_iou_0', 0.097), ('cardinality_error_0', 3.773), ('loss_caption_0', 2.843), ('loss_caption', 2.843), ('total_loss', 13.999)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 134000 (epoch 13),
loss = OrderedDict([('loss_ce', 0.273), ('loss_counter', 0.108), ('loss_bbox', 0.065), ('loss_giou', 0.171), ('loss_self_iou', 0.101), ('cardinality_error', 3.743), ('loss_ce_0', 0.276), ('loss_counter_0', 0.107), ('loss_bbox_0', 0.067), ('loss_giou_0', 0.179), ('loss_self_iou_0', 0.101), ('cardinality_error_0', 3.743), ('loss_caption_0', 2.786), ('loss_caption', 2.787), ('total_loss', 13.756)]),
time/iter = 0.188, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 135000 (epoch 13),
loss = OrderedDict([('loss_ce', 0.28), ('loss_counter', 0.115), ('loss_bbox', 0.061), ('loss_giou', 0.168), ('loss_self_iou', 0.096), ('cardinality_error', 3.794), ('loss_ce_0', 0.281), ('loss_counter_0', 0.115), ('loss_bbox_0', 0.064), ('loss_giou_0', 0.177), ('loss_self_iou_0', 0.097), ('cardinality_error_0', 3.794), ('loss_caption_0', 2.785), ('loss_caption', 2.784), ('total_loss', 13.759)]),
time/iter = 0.190, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 136000 (epoch 13),
loss = OrderedDict([('loss_ce', 0.279), ('loss_counter', 0.106), ('loss_bbox', 0.065), ('loss_giou', 0.168), ('loss_self_iou', 0.092), ('cardinality_error', 3.653), ('loss_ce_0', 0.279), ('loss_counter_0', 0.105), ('loss_bbox_0', 0.067), ('loss_giou_0', 0.175), ('loss_self_iou_0', 0.093), ('cardinality_error_0', 3.653), ('loss_caption_0', 2.828), ('loss_caption', 2.834), ('total_loss', 13.919)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 137000 (epoch 13),
loss = OrderedDict([('loss_ce', 0.279), ('loss_counter', 0.105), ('loss_bbox', 0.065), ('loss_giou', 0.173), ('loss_self_iou', 0.099), ('cardinality_error', 3.654), ('loss_ce_0', 0.281), ('loss_counter_0', 0.105), ('loss_bbox_0', 0.067), ('loss_giou_0', 0.179), ('loss_self_iou_0', 0.1), ('cardinality_error_0', 3.654), ('loss_caption_0', 2.79), ('loss_caption', 2.799), ('total_loss', 13.806)]),
time/iter = 0.190, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 138000 (epoch 13),
loss = OrderedDict([('loss_ce', 0.278), ('loss_counter', 0.109), ('loss_bbox', 0.064), ('loss_giou', 0.171), ('loss_self_iou', 0.095), ('cardinality_error', 3.714), ('loss_ce_0', 0.28), ('loss_counter_0', 0.108), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.178), ('loss_self_iou_0', 0.095), ('cardinality_error_0', 3.714), ('loss_caption_0', 2.835), ('loss_caption', 2.828), ('total_loss', 13.945)]),
time/iter = 0.188, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 139000 (epoch 13),
loss = OrderedDict([('loss_ce', 0.281), ('loss_counter', 0.115), ('loss_bbox', 0.062), ('loss_giou', 0.167), ('loss_self_iou', 0.098), ('cardinality_error', 3.813), ('loss_ce_0', 0.283), ('loss_counter_0', 0.114), ('loss_bbox_0', 0.064), ('loss_giou_0', 0.175), ('loss_self_iou_0', 0.099), ('cardinality_error_0', 3.813), ('loss_caption_0', 2.83), ('loss_caption', 2.828), ('total_loss', 13.924)]),
time/iter = 0.186, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 140000 (epoch 13),
loss = OrderedDict([('loss_ce', 0.277), ('loss_counter', 0.107), ('loss_bbox', 0.063), ('loss_giou', 0.171), ('loss_self_iou', 0.09), ('cardinality_error', 3.664), ('loss_ce_0', 0.28), ('loss_counter_0', 0.107), ('loss_bbox_0', 0.064), ('loss_giou_0', 0.178), ('loss_self_iou_0', 0.091), ('cardinality_error_0', 3.664), ('loss_caption_0', 2.821), ('loss_caption', 2.823), ('total_loss', 13.905)]),
time/iter = 0.191, bad_vid = 0.000
Validation results of iter 140126:
Bleu_1:0.16683698969676453
Bleu_2:0.09036855967772307
Bleu_3:0.047484441130632896
Bleu_4:0.023876859658376735
METEOR:0.08814626862844692
ROUGE_L:0.16473003568483396
CIDEr:0.3189568758512915
Recall:0.5281546209817979
Precision:0.5704333604501349
soda_c:0.057417105431783064
para_Bleu_1:0.4580706340663244
para_Bleu_2:0.27372623489326064
para_Bleu_3:0.16745128920972313
para_Bleu_4:0.10550306643408856
para_METEOR:0.16656454278617736
para_ROUGE_L:0.31631873012989425
para_CIDEr:0.19724321819057877
overall score of iter 140126: 0.46931082741084473
Save model at iter 140126 to /mnt/data/pjlab-3090-sport/wuhao/logs/dibs/howto-anet_anet_ori_pbox(similarity_op_order_v2)_CLIP/similarity_op_order_v2_topf30_iter3_r1_th1_refine_aug(8,0.02)_top3_2stage_ins_cap_topk_mil_coef0_noFocal_seq2-ft(mix)-gt_percent-1.0_2/model-last.pth.
Save info to info.json
ID seq2-ft(mix)-gt_percent-1.0 iter 141000 (epoch 14),
loss = OrderedDict([('loss_ce', 0.268), ('loss_counter', 0.108), ('loss_bbox', 0.066), ('loss_giou', 0.171), ('loss_self_iou', 0.106), ('cardinality_error', 3.774), ('loss_ce_0', 0.27), ('loss_counter_0', 0.108), ('loss_bbox_0', 0.067), ('loss_giou_0', 0.177), ('loss_self_iou_0', 0.108), ('cardinality_error_0', 3.774), ('loss_caption_0', 2.75), ('loss_caption', 2.748), ('total_loss', 13.572)]),
time/iter = 0.739, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 142000 (epoch 14),
loss = OrderedDict([('loss_ce', 0.27), ('loss_counter', 0.109), ('loss_bbox', 0.062), ('loss_giou', 0.173), ('loss_self_iou', 0.091), ('cardinality_error', 3.797), ('loss_ce_0', 0.272), ('loss_counter_0', 0.108), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.181), ('loss_self_iou_0', 0.091), ('cardinality_error_0', 3.797), ('loss_caption_0', 2.72), ('loss_caption', 2.722), ('total_loss', 13.492)]),
time/iter = 0.186, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 143000 (epoch 14),
loss = OrderedDict([('loss_ce', 0.265), ('loss_counter', 0.1), ('loss_bbox', 0.063), ('loss_giou', 0.162), ('loss_self_iou', 0.095), ('cardinality_error', 3.637), ('loss_ce_0', 0.268), ('loss_counter_0', 0.1), ('loss_bbox_0', 0.066), ('loss_giou_0', 0.171), ('loss_self_iou_0', 0.096), ('cardinality_error_0', 3.637), ('loss_caption_0', 2.782), ('loss_caption', 2.782), ('total_loss', 13.626)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 144000 (epoch 14),
loss = OrderedDict([('loss_ce', 0.27), ('loss_counter', 0.112), ('loss_bbox', 0.062), ('loss_giou', 0.172), ('loss_self_iou', 0.094), ('cardinality_error', 3.831), ('loss_ce_0', 0.273), ('loss_counter_0', 0.112), ('loss_bbox_0', 0.064), ('loss_giou_0', 0.18), ('loss_self_iou_0', 0.096), ('cardinality_error_0', 3.831), ('loss_caption_0', 2.793), ('loss_caption', 2.79), ('total_loss', 13.773)]),
time/iter = 0.189, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 145000 (epoch 14),
loss = OrderedDict([('loss_ce', 0.269), ('loss_counter', 0.101), ('loss_bbox', 0.061), ('loss_giou', 0.16), ('loss_self_iou', 0.093), ('cardinality_error', 3.665), ('loss_ce_0', 0.273), ('loss_counter_0', 0.101), ('loss_bbox_0', 0.063), ('loss_giou_0', 0.168), ('loss_self_iou_0', 0.093), ('cardinality_error_0', 3.665), ('loss_caption_0', 2.762), ('loss_caption', 2.767), ('total_loss', 13.554)]),
time/iter = 0.189, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 146000 (epoch 14),
loss = OrderedDict([('loss_ce', 0.275), ('loss_counter', 0.109), ('loss_bbox', 0.061), ('loss_giou', 0.164), ('loss_self_iou', 0.091), ('cardinality_error', 3.725), ('loss_ce_0', 0.276), ('loss_counter_0', 0.109), ('loss_bbox_0', 0.064), ('loss_giou_0', 0.172), ('loss_self_iou_0', 0.093), ('cardinality_error_0', 3.725), ('loss_caption_0', 2.813), ('loss_caption', 2.813), ('total_loss', 13.811)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 147000 (epoch 14),
loss = OrderedDict([('loss_ce', 0.272), ('loss_counter', 0.104), ('loss_bbox', 0.063), ('loss_giou', 0.171), ('loss_self_iou', 0.097), ('cardinality_error', 3.714), ('loss_ce_0', 0.273), ('loss_counter_0', 0.103), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.179), ('loss_self_iou_0', 0.098), ('cardinality_error_0', 3.714), ('loss_caption_0', 2.747), ('loss_caption', 2.745), ('total_loss', 13.578)]),
time/iter = 0.189, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 148000 (epoch 14),
loss = OrderedDict([('loss_ce', 0.271), ('loss_counter', 0.108), ('loss_bbox', 0.063), ('loss_giou', 0.168), ('loss_self_iou', 0.096), ('cardinality_error', 3.728), ('loss_ce_0', 0.274), ('loss_counter_0', 0.107), ('loss_bbox_0', 0.066), ('loss_giou_0', 0.177), ('loss_self_iou_0', 0.097), ('cardinality_error_0', 3.728), ('loss_caption_0', 2.843), ('loss_caption', 2.84), ('total_loss', 13.944)]),
time/iter = 0.192, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 149000 (epoch 14),
loss = OrderedDict([('loss_ce', 0.269), ('loss_counter', 0.108), ('loss_bbox', 0.066), ('loss_giou', 0.169), ('loss_self_iou', 0.098), ('cardinality_error', 3.799), ('loss_ce_0', 0.273), ('loss_counter_0', 0.109), ('loss_bbox_0', 0.068), ('loss_giou_0', 0.178), ('loss_self_iou_0', 0.099), ('cardinality_error_0', 3.799), ('loss_caption_0', 2.836), ('loss_caption', 2.836), ('total_loss', 13.926)]),
time/iter = 0.196, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 150000 (epoch 14),
loss = OrderedDict([('loss_ce', 0.27), ('loss_counter', 0.107), ('loss_bbox', 0.063), ('loss_giou', 0.169), ('loss_self_iou', 0.087), ('cardinality_error', 3.703), ('loss_ce_0', 0.272), ('loss_counter_0', 0.107), ('loss_bbox_0', 0.066), ('loss_giou_0', 0.176), ('loss_self_iou_0', 0.088), ('cardinality_error_0', 3.703), ('loss_caption_0', 2.806), ('loss_caption', 2.806), ('total_loss', 13.795)]),
time/iter = 0.193, bad_vid = 0.000
Validation results of iter 150135:
Bleu_1:0.16662144072598145
Bleu_2:0.08988753231411394
Bleu_3:0.04690847145308288
Bleu_4:0.023224274927987735
METEOR:0.08725158341768323
ROUGE_L:0.16364893754496343
CIDEr:0.32028824475030926
Recall:0.5260420675803493
Precision:0.5630584367161506
soda_c:0.057565785652999135
para_Bleu_1:0.46764194087144684
para_Bleu_2:0.2801629240374498
para_Bleu_3:0.1713033186995987
para_Bleu_4:0.10750827268624512
para_METEOR:0.16742715934059368
para_ROUGE_L:0.31858424377772926
para_CIDEr:0.2089956210595351
overall score of iter 150135: 0.4839310530863739
Save model at iter 150135 to /mnt/data/pjlab-3090-sport/wuhao/logs/dibs/howto-anet_anet_ori_pbox(similarity_op_order_v2)_CLIP/similarity_op_order_v2_topf30_iter3_r1_th1_refine_aug(8,0.02)_top3_2stage_ins_cap_topk_mil_coef0_noFocal_seq2-ft(mix)-gt_percent-1.0_2/model-last.pth.
Save Best-model at iter 150135 to checkpoint file.
Save info to info.json
ID seq2-ft(mix)-gt_percent-1.0 iter 151000 (epoch 15),
loss = OrderedDict([('loss_ce', 0.264), ('loss_counter', 0.101), ('loss_bbox', 0.063), ('loss_giou', 0.163), ('loss_self_iou', 0.097), ('cardinality_error', 3.645), ('loss_ce_0', 0.266), ('loss_counter_0', 0.101), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.171), ('loss_self_iou_0', 0.099), ('cardinality_error_0', 3.645), ('loss_caption_0', 2.762), ('loss_caption', 2.759), ('total_loss', 13.537)]),
time/iter = 0.737, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 152000 (epoch 15),
loss = OrderedDict([('loss_ce', 0.265), ('loss_counter', 0.103), ('loss_bbox', 0.06), ('loss_giou', 0.166), ('loss_self_iou', 0.087), ('cardinality_error', 3.722), ('loss_ce_0', 0.269), ('loss_counter_0', 0.103), ('loss_bbox_0', 0.063), ('loss_giou_0', 0.175), ('loss_self_iou_0', 0.087), ('cardinality_error_0', 3.722), ('loss_caption_0', 2.762), ('loss_caption', 2.766), ('total_loss', 13.59)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 153000 (epoch 15),
loss = OrderedDict([('loss_ce', 0.264), ('loss_counter', 0.111), ('loss_bbox', 0.062), ('loss_giou', 0.168), ('loss_self_iou', 0.083), ('cardinality_error', 3.813), ('loss_ce_0', 0.267), ('loss_counter_0', 0.111), ('loss_bbox_0', 0.064), ('loss_giou_0', 0.177), ('loss_self_iou_0', 0.085), ('cardinality_error_0', 3.813), ('loss_caption_0', 2.777), ('loss_caption', 2.778), ('total_loss', 13.663)]),
time/iter = 0.193, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 154000 (epoch 15),
loss = OrderedDict([('loss_ce', 0.268), ('loss_counter', 0.106), ('loss_bbox', 0.061), ('loss_giou', 0.168), ('loss_self_iou', 0.092), ('cardinality_error', 3.769), ('loss_ce_0', 0.272), ('loss_counter_0', 0.105), ('loss_bbox_0', 0.064), ('loss_giou_0', 0.178), ('loss_self_iou_0', 0.092), ('cardinality_error_0', 3.769), ('loss_caption_0', 2.787), ('loss_caption', 2.787), ('total_loss', 13.717)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 155000 (epoch 15),
loss = OrderedDict([('loss_ce', 0.264), ('loss_counter', 0.104), ('loss_bbox', 0.063), ('loss_giou', 0.169), ('loss_self_iou', 0.09), ('cardinality_error', 3.714), ('loss_ce_0', 0.267), ('loss_counter_0', 0.104), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.179), ('loss_self_iou_0', 0.091), ('cardinality_error_0', 3.714), ('loss_caption_0', 2.758), ('loss_caption', 2.76), ('total_loss', 13.593)]),
time/iter = 0.190, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 156000 (epoch 15),
loss = OrderedDict([('loss_ce', 0.265), ('loss_counter', 0.106), ('loss_bbox', 0.064), ('loss_giou', 0.167), ('loss_self_iou', 0.102), ('cardinality_error', 3.675), ('loss_ce_0', 0.269), ('loss_counter_0', 0.106), ('loss_bbox_0', 0.066), ('loss_giou_0', 0.174), ('loss_self_iou_0', 0.102), ('cardinality_error_0', 3.675), ('loss_caption_0', 2.741), ('loss_caption', 2.742), ('total_loss', 13.504)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 157000 (epoch 15),
loss = OrderedDict([('loss_ce', 0.267), ('loss_counter', 0.104), ('loss_bbox', 0.065), ('loss_giou', 0.167), ('loss_self_iou', 0.103), ('cardinality_error', 3.722), ('loss_ce_0', 0.268), ('loss_counter_0', 0.104), ('loss_bbox_0', 0.068), ('loss_giou_0', 0.176), ('loss_self_iou_0', 0.105), ('cardinality_error_0', 3.722), ('loss_caption_0', 2.777), ('loss_caption', 2.783), ('total_loss', 13.668)]),
time/iter = 0.188, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 158000 (epoch 15),
loss = OrderedDict([('loss_ce', 0.266), ('loss_counter', 0.106), ('loss_bbox', 0.062), ('loss_giou', 0.164), ('loss_self_iou', 0.099), ('cardinality_error', 3.758), ('loss_ce_0', 0.27), ('loss_counter_0', 0.106), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.173), ('loss_self_iou_0', 0.099), ('cardinality_error_0', 3.758), ('loss_caption_0', 2.815), ('loss_caption', 2.817), ('total_loss', 13.789)]),
time/iter = 0.192, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 159000 (epoch 15),
loss = OrderedDict([('loss_ce', 0.272), ('loss_counter', 0.108), ('loss_bbox', 0.062), ('loss_giou', 0.169), ('loss_self_iou', 0.098), ('cardinality_error', 3.729), ('loss_ce_0', 0.275), ('loss_counter_0', 0.108), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.177), ('loss_self_iou_0', 0.1), ('cardinality_error_0', 3.729), ('loss_caption_0', 2.783), ('loss_caption', 2.785), ('total_loss', 13.721)]),
time/iter = 0.194, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 160000 (epoch 15),
loss = OrderedDict([('loss_ce', 0.269), ('loss_counter', 0.109), ('loss_bbox', 0.063), ('loss_giou', 0.166), ('loss_self_iou', 0.098), ('cardinality_error', 3.816), ('loss_ce_0', 0.271), ('loss_counter_0', 0.109), ('loss_bbox_0', 0.066), ('loss_giou_0', 0.176), ('loss_self_iou_0', 0.098), ('cardinality_error_0', 3.816), ('loss_caption_0', 2.78), ('loss_caption', 2.784), ('total_loss', 13.686)]),
time/iter = 0.196, bad_vid = 0.000
Validation results of iter 160144:
Bleu_1:0.16754398447821903
Bleu_2:0.08978801866243748
Bleu_3:0.046077601805781236
Bleu_4:0.02215727819941335
METEOR:0.08650894641812401
ROUGE_L:0.16425299709373153
CIDEr:0.3192637628790779
Recall:0.5308598805776927
Precision:0.5705477594739302
soda_c:0.059035206979637336
para_Bleu_1:0.4722129873397206
para_Bleu_2:0.2843271953295457
para_Bleu_3:0.17433620623201318
para_Bleu_4:0.10943737200004257
para_METEOR:0.16524483023272712
para_ROUGE_L:0.3180351825656492
para_CIDEr:0.2139382514781602
overall score of iter 160144: 0.4886204537109299
Save model at iter 160144 to /mnt/data/pjlab-3090-sport/wuhao/logs/dibs/howto-anet_anet_ori_pbox(similarity_op_order_v2)_CLIP/similarity_op_order_v2_topf30_iter3_r1_th1_refine_aug(8,0.02)_top3_2stage_ins_cap_topk_mil_coef0_noFocal_seq2-ft(mix)-gt_percent-1.0_2/model-last.pth.
Save Best-model at iter 160144 to checkpoint file.
Save info to info.json
ID seq2-ft(mix)-gt_percent-1.0 iter 161000 (epoch 16),
loss = OrderedDict([('loss_ce', 0.26), ('loss_counter', 0.103), ('loss_bbox', 0.061), ('loss_giou', 0.163), ('loss_self_iou', 0.097), ('cardinality_error', 3.695), ('loss_ce_0', 0.263), ('loss_counter_0', 0.103), ('loss_bbox_0', 0.063), ('loss_giou_0', 0.171), ('loss_self_iou_0', 0.098), ('cardinality_error_0', 3.695), ('loss_caption_0', 2.766), ('loss_caption', 2.768), ('total_loss', 13.553)]),
time/iter = 0.749, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 162000 (epoch 16),
loss = OrderedDict([('loss_ce', 0.262), ('loss_counter', 0.103), ('loss_bbox', 0.063), ('loss_giou', 0.164), ('loss_self_iou', 0.091), ('cardinality_error', 3.694), ('loss_ce_0', 0.266), ('loss_counter_0', 0.103), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.174), ('loss_self_iou_0', 0.092), ('cardinality_error_0', 3.694), ('loss_caption_0', 2.768), ('loss_caption', 2.764), ('total_loss', 13.573)]),
time/iter = 0.188, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 163000 (epoch 16),
loss = OrderedDict([('loss_ce', 0.262), ('loss_counter', 0.105), ('loss_bbox', 0.064), ('loss_giou', 0.173), ('loss_self_iou', 0.097), ('cardinality_error', 3.769), ('loss_ce_0', 0.266), ('loss_counter_0', 0.104), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.179), ('loss_self_iou_0', 0.098), ('cardinality_error_0', 3.769), ('loss_caption_0', 2.765), ('loss_caption', 2.766), ('total_loss', 13.63)]),
time/iter = 0.189, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 164000 (epoch 16),
loss = OrderedDict([('loss_ce', 0.265), ('loss_counter', 0.11), ('loss_bbox', 0.061), ('loss_giou', 0.164), ('loss_self_iou', 0.092), ('cardinality_error', 3.774), ('loss_ce_0', 0.269), ('loss_counter_0', 0.11), ('loss_bbox_0', 0.063), ('loss_giou_0', 0.173), ('loss_self_iou_0', 0.092), ('cardinality_error_0', 3.774), ('loss_caption_0', 2.772), ('loss_caption', 2.776), ('total_loss', 13.625)]),
time/iter = 0.190, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 165000 (epoch 16),
loss = OrderedDict([('loss_ce', 0.264), ('loss_counter', 0.102), ('loss_bbox', 0.063), ('loss_giou', 0.164), ('loss_self_iou', 0.092), ('cardinality_error', 3.699), ('loss_ce_0', 0.267), ('loss_counter_0', 0.102), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.173), ('loss_self_iou_0', 0.092), ('cardinality_error_0', 3.699), ('loss_caption_0', 2.711), ('loss_caption', 2.716), ('total_loss', 13.368)]),
time/iter = 0.187, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 166000 (epoch 16),
loss = OrderedDict([('loss_ce', 0.264), ('loss_counter', 0.105), ('loss_bbox', 0.061), ('loss_giou', 0.163), ('loss_self_iou', 0.094), ('cardinality_error', 3.72), ('loss_ce_0', 0.268), ('loss_counter_0', 0.105), ('loss_bbox_0', 0.063), ('loss_giou_0', 0.174), ('loss_self_iou_0', 0.095), ('cardinality_error_0', 3.72), ('loss_caption_0', 2.754), ('loss_caption', 2.755), ('total_loss', 13.534)]),
time/iter = 0.193, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 167000 (epoch 16),
loss = OrderedDict([('loss_ce', 0.261), ('loss_counter', 0.101), ('loss_bbox', 0.062), ('loss_giou', 0.168), ('loss_self_iou', 0.095), ('cardinality_error', 3.712), ('loss_ce_0', 0.266), ('loss_counter_0', 0.1), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.176), ('loss_self_iou_0', 0.095), ('cardinality_error_0', 3.712), ('loss_caption_0', 2.771), ('loss_caption', 2.772), ('total_loss', 13.617)]),
time/iter = 0.192, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 168000 (epoch 16),
loss = OrderedDict([('loss_ce', 0.265), ('loss_counter', 0.108), ('loss_bbox', 0.062), ('loss_giou', 0.168), ('loss_self_iou', 0.09), ('cardinality_error', 3.816), ('loss_ce_0', 0.269), ('loss_counter_0', 0.107), ('loss_bbox_0', 0.064), ('loss_giou_0', 0.177), ('loss_self_iou_0', 0.092), ('cardinality_error_0', 3.816), ('loss_caption_0', 2.814), ('loss_caption', 2.82), ('total_loss', 13.826)]),
time/iter = 0.193, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 169000 (epoch 16),
loss = OrderedDict([('loss_ce', 0.258), ('loss_counter', 0.106), ('loss_bbox', 0.064), ('loss_giou', 0.166), ('loss_self_iou', 0.106), ('cardinality_error', 3.697), ('loss_ce_0', 0.261), ('loss_counter_0', 0.106), ('loss_bbox_0', 0.067), ('loss_giou_0', 0.176), ('loss_self_iou_0', 0.107), ('cardinality_error_0', 3.697), ('loss_caption_0', 2.769), ('loss_caption', 2.775), ('total_loss', 13.598)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 170000 (epoch 16),
loss = OrderedDict([('loss_ce', 0.268), ('loss_counter', 0.105), ('loss_bbox', 0.062), ('loss_giou', 0.165), ('loss_self_iou', 0.093), ('cardinality_error', 3.799), ('loss_ce_0', 0.272), ('loss_counter_0', 0.106), ('loss_bbox_0', 0.064), ('loss_giou_0', 0.174), ('loss_self_iou_0', 0.094), ('cardinality_error_0', 3.799), ('loss_caption_0', 2.794), ('loss_caption', 2.798), ('total_loss', 13.727)]),
time/iter = 0.191, bad_vid = 0.000
Validation results of iter 170153:
Bleu_1:0.16584280243722227
Bleu_2:0.08889969905794425
Bleu_3:0.04569298286173284
Bleu_4:0.021992960199339176
METEOR:0.08570833880397384
ROUGE_L:0.16234979503724006
CIDEr:0.3170462149966731
Recall:0.5273397281824633
Precision:0.5648989898989865
soda_c:0.058539462474976364
para_Bleu_1:0.4735378044184376
para_Bleu_2:0.2855599966961999
para_Bleu_3:0.17485842077678387
para_Bleu_4:0.10998333079246524
para_METEOR:0.16580782598840993
para_ROUGE_L:0.3184105968751349
para_CIDEr:0.2144083270960459
overall score of iter 170153: 0.4901994838769211
Save model at iter 170153 to /mnt/data/pjlab-3090-sport/wuhao/logs/dibs/howto-anet_anet_ori_pbox(similarity_op_order_v2)_CLIP/similarity_op_order_v2_topf30_iter3_r1_th1_refine_aug(8,0.02)_top3_2stage_ins_cap_topk_mil_coef0_noFocal_seq2-ft(mix)-gt_percent-1.0_2/model-last.pth.
Save Best-model at iter 170153 to checkpoint file.
Save info to info.json
ID seq2-ft(mix)-gt_percent-1.0 iter 171000 (epoch 17),
loss = OrderedDict([('loss_ce', 0.256), ('loss_counter', 0.101), ('loss_bbox', 0.062), ('loss_giou', 0.161), ('loss_self_iou', 0.094), ('cardinality_error', 3.694), ('loss_ce_0', 0.261), ('loss_counter_0', 0.101), ('loss_bbox_0', 0.064), ('loss_giou_0', 0.169), ('loss_self_iou_0', 0.093), ('cardinality_error_0', 3.694), ('loss_caption_0', 2.772), ('loss_caption', 2.77), ('total_loss', 13.544)]),
time/iter = 0.745, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 172000 (epoch 17),
loss = OrderedDict([('loss_ce', 0.258), ('loss_counter', 0.1), ('loss_bbox', 0.063), ('loss_giou', 0.165), ('loss_self_iou', 0.096), ('cardinality_error', 3.667), ('loss_ce_0', 0.262), ('loss_counter_0', 0.1), ('loss_bbox_0', 0.066), ('loss_giou_0', 0.175), ('loss_self_iou_0', 0.095), ('cardinality_error_0', 3.667), ('loss_caption_0', 2.741), ('loss_caption', 2.743), ('total_loss', 13.47)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 173000 (epoch 17),
loss = OrderedDict([('loss_ce', 0.258), ('loss_counter', 0.104), ('loss_bbox', 0.062), ('loss_giou', 0.165), ('loss_self_iou', 0.09), ('cardinality_error', 3.753), ('loss_ce_0', 0.261), ('loss_counter_0', 0.104), ('loss_bbox_0', 0.064), ('loss_giou_0', 0.175), ('loss_self_iou_0', 0.091), ('cardinality_error_0', 3.753), ('loss_caption_0', 2.786), ('loss_caption', 2.785), ('total_loss', 13.646)]),
time/iter = 0.193, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 174000 (epoch 17),
loss = OrderedDict([('loss_ce', 0.259), ('loss_counter', 0.107), ('loss_bbox', 0.06), ('loss_giou', 0.166), ('loss_self_iou', 0.094), ('cardinality_error', 3.832), ('loss_ce_0', 0.261), ('loss_counter_0', 0.107), ('loss_bbox_0', 0.064), ('loss_giou_0', 0.177), ('loss_self_iou_0', 0.096), ('cardinality_error_0', 3.832), ('loss_caption_0', 2.733), ('loss_caption', 2.738), ('total_loss', 13.457)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 175000 (epoch 17),
loss = OrderedDict([('loss_ce', 0.255), ('loss_counter', 0.103), ('loss_bbox', 0.06), ('loss_giou', 0.163), ('loss_self_iou', 0.098), ('cardinality_error', 3.731), ('loss_ce_0', 0.259), ('loss_counter_0', 0.103), ('loss_bbox_0', 0.062), ('loss_giou_0', 0.173), ('loss_self_iou_0', 0.098), ('cardinality_error_0', 3.731), ('loss_caption_0', 2.745), ('loss_caption', 2.744), ('total_loss', 13.454)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 176000 (epoch 17),
loss = OrderedDict([('loss_ce', 0.261), ('loss_counter', 0.103), ('loss_bbox', 0.06), ('loss_giou', 0.164), ('loss_self_iou', 0.095), ('cardinality_error', 3.795), ('loss_ce_0', 0.264), ('loss_counter_0', 0.103), ('loss_bbox_0', 0.064), ('loss_giou_0', 0.176), ('loss_self_iou_0', 0.095), ('cardinality_error_0', 3.795), ('loss_caption_0', 2.761), ('loss_caption', 2.77), ('total_loss', 13.575)]),
time/iter = 0.192, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 177000 (epoch 17),
loss = OrderedDict([('loss_ce', 0.255), ('loss_counter', 0.1), ('loss_bbox', 0.063), ('loss_giou', 0.161), ('loss_self_iou', 0.096), ('cardinality_error', 3.652), ('loss_ce_0', 0.261), ('loss_counter_0', 0.1), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.169), ('loss_self_iou_0', 0.098), ('cardinality_error_0', 3.652), ('loss_caption_0', 2.743), ('loss_caption', 2.745), ('total_loss', 13.43)]),
time/iter = 0.190, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 178000 (epoch 17),
loss = OrderedDict([('loss_ce', 0.255), ('loss_counter', 0.103), ('loss_bbox', 0.063), ('loss_giou', 0.164), ('loss_self_iou', 0.103), ('cardinality_error', 3.664), ('loss_ce_0', 0.26), ('loss_counter_0', 0.102), ('loss_bbox_0', 0.066), ('loss_giou_0', 0.175), ('loss_self_iou_0', 0.104), ('cardinality_error_0', 3.664), ('loss_caption_0', 2.682), ('loss_caption', 2.68), ('total_loss', 13.211)]),
time/iter = 0.188, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 179000 (epoch 17),
loss = OrderedDict([('loss_ce', 0.261), ('loss_counter', 0.105), ('loss_bbox', 0.06), ('loss_giou', 0.164), ('loss_self_iou', 0.09), ('cardinality_error', 3.825), ('loss_ce_0', 0.266), ('loss_counter_0', 0.104), ('loss_bbox_0', 0.063), ('loss_giou_0', 0.173), ('loss_self_iou_0', 0.091), ('cardinality_error_0', 3.825), ('loss_caption_0', 2.788), ('loss_caption', 2.796), ('total_loss', 13.671)]),
time/iter = 0.193, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 180000 (epoch 17),
loss = OrderedDict([('loss_ce', 0.255), ('loss_counter', 0.102), ('loss_bbox', 0.064), ('loss_giou', 0.166), ('loss_self_iou', 0.093), ('cardinality_error', 3.729), ('loss_ce_0', 0.261), ('loss_counter_0', 0.102), ('loss_bbox_0', 0.066), ('loss_giou_0', 0.175), ('loss_self_iou_0', 0.094), ('cardinality_error_0', 3.729), ('loss_caption_0', 2.781), ('loss_caption', 2.775), ('total_loss', 13.608)]),
time/iter = 0.192, bad_vid = 0.000
Validation results of iter 180162:
Bleu_1:0.16720622564646215
Bleu_2:0.08946643461131876
Bleu_3:0.04568137095423273
Bleu_4:0.022039722503534608
METEOR:0.08588931176535387
ROUGE_L:0.16315869782389542
CIDEr:0.32099741016990446
Recall:0.5265047853249455
Precision:0.5647345942647923
soda_c:0.05847424883094643
para_Bleu_1:0.47508155945278135
para_Bleu_2:0.2858233856765029
para_Bleu_3:0.17499503512152859
para_Bleu_4:0.11002968407978216
para_METEOR:0.16541373751181562
para_ROUGE_L:0.3190110890037882
para_CIDEr:0.21421557986951392
overall score of iter 180162: 0.4896590014611117
Save model at iter 180162 to /mnt/data/pjlab-3090-sport/wuhao/logs/dibs/howto-anet_anet_ori_pbox(similarity_op_order_v2)_CLIP/similarity_op_order_v2_topf30_iter3_r1_th1_refine_aug(8,0.02)_top3_2stage_ins_cap_topk_mil_coef0_noFocal_seq2-ft(mix)-gt_percent-1.0_2/model-last.pth.
Save info to info.json
ID seq2-ft(mix)-gt_percent-1.0 iter 181000 (epoch 18),
loss = OrderedDict([('loss_ce', 0.256), ('loss_counter', 0.102), ('loss_bbox', 0.061), ('loss_giou', 0.163), ('loss_self_iou', 0.094), ('cardinality_error', 3.781), ('loss_ce_0', 0.261), ('loss_counter_0', 0.102), ('loss_bbox_0', 0.063), ('loss_giou_0', 0.172), ('loss_self_iou_0', 0.094), ('cardinality_error_0', 3.781), ('loss_caption_0', 2.743), ('loss_caption', 2.746), ('total_loss', 13.452)]),
time/iter = 0.750, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 182000 (epoch 18),
loss = OrderedDict([('loss_ce', 0.255), ('loss_counter', 0.101), ('loss_bbox', 0.062), ('loss_giou', 0.164), ('loss_self_iou', 0.1), ('cardinality_error', 3.726), ('loss_ce_0', 0.26), ('loss_counter_0', 0.101), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.174), ('loss_self_iou_0', 0.1), ('cardinality_error_0', 3.726), ('loss_caption_0', 2.748), ('loss_caption', 2.746), ('total_loss', 13.472)]),
time/iter = 0.189, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 183000 (epoch 18),
loss = OrderedDict([('loss_ce', 0.256), ('loss_counter', 0.102), ('loss_bbox', 0.061), ('loss_giou', 0.163), ('loss_self_iou', 0.097), ('cardinality_error', 3.722), ('loss_ce_0', 0.26), ('loss_counter_0', 0.102), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.174), ('loss_self_iou_0', 0.098), ('cardinality_error_0', 3.722), ('loss_caption_0', 2.729), ('loss_caption', 2.734), ('total_loss', 13.405)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 184000 (epoch 18),
loss = OrderedDict([('loss_ce', 0.253), ('loss_counter', 0.104), ('loss_bbox', 0.061), ('loss_giou', 0.161), ('loss_self_iou', 0.098), ('cardinality_error', 3.726), ('loss_ce_0', 0.257), ('loss_counter_0', 0.104), ('loss_bbox_0', 0.064), ('loss_giou_0', 0.17), ('loss_self_iou_0', 0.098), ('cardinality_error_0', 3.726), ('loss_caption_0', 2.783), ('loss_caption', 2.787), ('total_loss', 13.591)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 185000 (epoch 18),
loss = OrderedDict([('loss_ce', 0.255), ('loss_counter', 0.098), ('loss_bbox', 0.063), ('loss_giou', 0.165), ('loss_self_iou', 0.087), ('cardinality_error', 3.667), ('loss_ce_0', 0.26), ('loss_counter_0', 0.098), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.175), ('loss_self_iou_0', 0.088), ('cardinality_error_0', 3.667), ('loss_caption_0', 2.718), ('loss_caption', 2.716), ('total_loss', 13.354)]),
time/iter = 0.188, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 186000 (epoch 18),
loss = OrderedDict([('loss_ce', 0.254), ('loss_counter', 0.099), ('loss_bbox', 0.062), ('loss_giou', 0.166), ('loss_self_iou', 0.093), ('cardinality_error', 3.776), ('loss_ce_0', 0.259), ('loss_counter_0', 0.099), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.176), ('loss_self_iou_0', 0.094), ('cardinality_error_0', 3.776), ('loss_caption_0', 2.75), ('loss_caption', 2.75), ('total_loss', 13.494)]),
time/iter = 0.194, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 187000 (epoch 18),
loss = OrderedDict([('loss_ce', 0.258), ('loss_counter', 0.109), ('loss_bbox', 0.062), ('loss_giou', 0.165), ('loss_self_iou', 0.089), ('cardinality_error', 3.803), ('loss_ce_0', 0.264), ('loss_counter_0', 0.109), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.176), ('loss_self_iou_0', 0.09), ('cardinality_error_0', 3.803), ('loss_caption_0', 2.788), ('loss_caption', 2.791), ('total_loss', 13.678)]),
time/iter = 0.198, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 188000 (epoch 18),
loss = OrderedDict([('loss_ce', 0.253), ('loss_counter', 0.1), ('loss_bbox', 0.062), ('loss_giou', 0.163), ('loss_self_iou', 0.091), ('cardinality_error', 3.71), ('loss_ce_0', 0.259), ('loss_counter_0', 0.1), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.173), ('loss_self_iou_0', 0.092), ('cardinality_error_0', 3.71), ('loss_caption_0', 2.745), ('loss_caption', 2.743), ('total_loss', 13.444)]),
time/iter = 0.192, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 189000 (epoch 18),
loss = OrderedDict([('loss_ce', 0.25), ('loss_counter', 0.105), ('loss_bbox', 0.064), ('loss_giou', 0.165), ('loss_self_iou', 0.1), ('cardinality_error', 3.748), ('loss_ce_0', 0.256), ('loss_counter_0', 0.105), ('loss_bbox_0', 0.067), ('loss_giou_0', 0.175), ('loss_self_iou_0', 0.101), ('cardinality_error_0', 3.748), ('loss_caption_0', 2.751), ('loss_caption', 2.753), ('total_loss', 13.484)]),
time/iter = 0.188, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 190000 (epoch 18),
loss = OrderedDict([('loss_ce', 0.257), ('loss_counter', 0.104), ('loss_bbox', 0.06), ('loss_giou', 0.161), ('loss_self_iou', 0.098), ('cardinality_error', 3.742), ('loss_ce_0', 0.264), ('loss_counter_0', 0.104), ('loss_bbox_0', 0.063), ('loss_giou_0', 0.172), ('loss_self_iou_0', 0.099), ('cardinality_error_0', 3.742), ('loss_caption_0', 2.729), ('loss_caption', 2.73), ('total_loss', 13.395)]),
time/iter = 0.189, bad_vid = 0.000
Validation results of iter 190171:
Bleu_1:0.1662475028889873
Bleu_2:0.08895418147726737
Bleu_3:0.04559170272578064
Bleu_4:0.021869443641790748
METEOR:0.0853620749347768
ROUGE_L:0.16226693807975517
CIDEr:0.3203697867996399
Recall:0.5243080966273422
Precision:0.5592002237136435
soda_c:0.058066485957305666
para_Bleu_1:0.47302383939773723
para_Bleu_2:0.2848420020452884
para_Bleu_3:0.17477626094199183
para_Bleu_4:0.11005159892431456
para_METEOR:0.16474042555391544
para_ROUGE_L:0.31754161420686944
para_CIDEr:0.2082818020277855
overall score of iter 190171: 0.4830738265060155
Save model at iter 190171 to /mnt/data/pjlab-3090-sport/wuhao/logs/dibs/howto-anet_anet_ori_pbox(similarity_op_order_v2)_CLIP/similarity_op_order_v2_topf30_iter3_r1_th1_refine_aug(8,0.02)_top3_2stage_ins_cap_topk_mil_coef0_noFocal_seq2-ft(mix)-gt_percent-1.0_2/model-last.pth.
Save info to info.json
ID seq2-ft(mix)-gt_percent-1.0 iter 191000 (epoch 19),
loss = OrderedDict([('loss_ce', 0.251), ('loss_counter', 0.099), ('loss_bbox', 0.062), ('loss_giou', 0.167), ('loss_self_iou', 0.086), ('cardinality_error', 3.653), ('loss_ce_0', 0.257), ('loss_counter_0', 0.099), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.176), ('loss_self_iou_0', 0.087), ('cardinality_error_0', 3.653), ('loss_caption_0', 2.754), ('loss_caption', 2.752), ('total_loss', 13.501)]),
time/iter = 0.755, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 192000 (epoch 19),
loss = OrderedDict([('loss_ce', 0.252), ('loss_counter', 0.1), ('loss_bbox', 0.061), ('loss_giou', 0.164), ('loss_self_iou', 0.094), ('cardinality_error', 3.767), ('loss_ce_0', 0.258), ('loss_counter_0', 0.1), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.173), ('loss_self_iou_0', 0.097), ('cardinality_error_0', 3.767), ('loss_caption_0', 2.717), ('loss_caption', 2.72), ('total_loss', 13.343)]),
time/iter = 0.188, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 193000 (epoch 19),
loss = OrderedDict([('loss_ce', 0.25), ('loss_counter', 0.106), ('loss_bbox', 0.06), ('loss_giou', 0.164), ('loss_self_iou', 0.093), ('cardinality_error', 3.847), ('loss_ce_0', 0.256), ('loss_counter_0', 0.105), ('loss_bbox_0', 0.063), ('loss_giou_0', 0.174), ('loss_self_iou_0', 0.093), ('cardinality_error_0', 3.847), ('loss_caption_0', 2.754), ('loss_caption', 2.759), ('total_loss', 13.499)]),
time/iter = 0.190, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 194000 (epoch 19),
loss = OrderedDict([('loss_ce', 0.256), ('loss_counter', 0.102), ('loss_bbox', 0.061), ('loss_giou', 0.165), ('loss_self_iou', 0.097), ('cardinality_error', 3.775), ('loss_ce_0', 0.262), ('loss_counter_0', 0.102), ('loss_bbox_0', 0.063), ('loss_giou_0', 0.176), ('loss_self_iou_0', 0.098), ('cardinality_error_0', 3.775), ('loss_caption_0', 2.769), ('loss_caption', 2.772), ('total_loss', 13.587)]),
time/iter = 0.192, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 195000 (epoch 19),
loss = OrderedDict([('loss_ce', 0.257), ('loss_counter', 0.106), ('loss_bbox', 0.062), ('loss_giou', 0.165), ('loss_self_iou', 0.089), ('cardinality_error', 3.794), ('loss_ce_0', 0.261), ('loss_counter_0', 0.105), ('loss_bbox_0', 0.064), ('loss_giou_0', 0.175), ('loss_self_iou_0', 0.089), ('cardinality_error_0', 3.794), ('loss_caption_0', 2.751), ('loss_caption', 2.751), ('total_loss', 13.506)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 196000 (epoch 19),
loss = OrderedDict([('loss_ce', 0.251), ('loss_counter', 0.095), ('loss_bbox', 0.061), ('loss_giou', 0.162), ('loss_self_iou', 0.1), ('cardinality_error', 3.652), ('loss_ce_0', 0.258), ('loss_counter_0', 0.095), ('loss_bbox_0', 0.064), ('loss_giou_0', 0.172), ('loss_self_iou_0', 0.101), ('cardinality_error_0', 3.652), ('loss_caption_0', 2.743), ('loss_caption', 2.735), ('total_loss', 13.403)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 197000 (epoch 19),
loss = OrderedDict([('loss_ce', 0.251), ('loss_counter', 0.104), ('loss_bbox', 0.061), ('loss_giou', 0.162), ('loss_self_iou', 0.091), ('cardinality_error', 3.759), ('loss_ce_0', 0.258), ('loss_counter_0', 0.104), ('loss_bbox_0', 0.064), ('loss_giou_0', 0.171), ('loss_self_iou_0', 0.091), ('cardinality_error_0', 3.759), ('loss_caption_0', 2.74), ('loss_caption', 2.743), ('total_loss', 13.418)]),
time/iter = 0.191, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 198000 (epoch 19),
loss = OrderedDict([('loss_ce', 0.249), ('loss_counter', 0.098), ('loss_bbox', 0.062), ('loss_giou', 0.162), ('loss_self_iou', 0.092), ('cardinality_error', 3.664), ('loss_ce_0', 0.255), ('loss_counter_0', 0.098), ('loss_bbox_0', 0.064), ('loss_giou_0', 0.171), ('loss_self_iou_0', 0.092), ('cardinality_error_0', 3.664), ('loss_caption_0', 2.718), ('loss_caption', 2.72), ('total_loss', 13.31)]),
time/iter = 0.190, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 199000 (epoch 19),
loss = OrderedDict([('loss_ce', 0.252), ('loss_counter', 0.101), ('loss_bbox', 0.062), ('loss_giou', 0.162), ('loss_self_iou', 0.101), ('cardinality_error', 3.736), ('loss_ce_0', 0.257), ('loss_counter_0', 0.101), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.173), ('loss_self_iou_0', 0.102), ('cardinality_error_0', 3.736), ('loss_caption_0', 2.759), ('loss_caption', 2.76), ('total_loss', 13.502)]),
time/iter = 0.193, bad_vid = 0.000
ID seq2-ft(mix)-gt_percent-1.0 iter 200000 (epoch 19),
loss = OrderedDict([('loss_ce', 0.253), ('loss_counter', 0.102), ('loss_bbox', 0.061), ('loss_giou', 0.159), ('loss_self_iou', 0.098), ('cardinality_error', 3.701), ('loss_ce_0', 0.259), ('loss_counter_0', 0.102), ('loss_bbox_0', 0.065), ('loss_giou_0', 0.17), ('loss_self_iou_0', 0.098), ('cardinality_error_0', 3.701), ('loss_caption_0', 2.766), ('loss_caption', 2.771), ('total_loss', 13.518)]),
time/iter = 0.190, bad_vid = 0.000
Validation results of iter 200180:
Bleu_1:0.16600244771432068
Bleu_2:0.08859363359362551
Bleu_3:0.045174799285766926
Bleu_4:0.021453706973694267
METEOR:0.08469975853590762
ROUGE_L:0.1615333099598977
CIDEr:0.3178372173219055
Recall:0.5270524681293403
Precision:0.5612365263371945
soda_c:0.05852570981425518
para_Bleu_1:0.47641872729084495
para_Bleu_2:0.28679556025023933
para_Bleu_3:0.1757988669447671
para_Bleu_4:0.11061748158923715
para_METEOR:0.1647238014039032
para_ROUGE_L:0.3182336912910021
para_CIDEr:0.21852415031403352
overall score of iter 200180: 0.4938654333071738
Save model at iter 200180 to /mnt/data/pjlab-3090-sport/wuhao/logs/dibs/howto-anet_anet_ori_pbox(similarity_op_order_v2)_CLIP/similarity_op_order_v2_topf30_iter3_r1_th1_refine_aug(8,0.02)_top3_2stage_ins_cap_topk_mil_coef0_noFocal_seq2-ft(mix)-gt_percent-1.0_2/model-last.pth.
Save Best-model at iter 200180 to checkpoint file.
Save info to info.json
Best epoch: 10
Best Model Performance:
Bleu_1:0.1671778590456048
Bleu_2:0.09077014613023152
Bleu_3:0.0476684747303012
Bleu_4:0.02445564298599047
METEOR:0.08933235383587503
ROUGE_L:0.1654660162888944
CIDEr:0.31886265111118334
Recall:0.5314017615268335
Precision:0.5831469052945512
soda_c:0.05853263249839839
para_Bleu_1:0.46544090189732323
para_Bleu_2:0.2789325258737778
para_Bleu_3:0.17172911957785325
para_Bleu_4:0.10903514181091935
para_METEOR:0.16550159188298816
para_ROUGE_L:0.3181118223429575
para_CIDEr:0.2056618808195008
avg_proposal_number:-1
Best Overall Score epoch10: 1.5812763042668414
|