detr_finetuned_30

This model is a fine-tuned version of facebook/detr-resnet-50 on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 1.1249
  • Map: 0.2684
  • Map 50: 0.5212
  • Map 75: 0.2578
  • Map Small: 0.22
  • Map Medium: 0.377
  • Map Large: 0.395
  • Mar 1: 0.1497
  • Mar 10: 0.3903
  • Mar 100: 0.4383
  • Mar Small: 0.398
  • Mar Medium: 0.5486
  • Mar Large: 0.6314
  • Map Basketball: 0.0431
  • Mar 100 Basketball: 0.147
  • Map Player: 0.3203
  • Mar 100 Player: 0.5743
  • Map Referee: 0.4419
  • Mar 100 Referee: 0.5936

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 5e-05
  • train_batch_size: 16
  • eval_batch_size: 8
  • seed: 42
  • optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
  • lr_scheduler_type: cosine
  • num_epochs: 30

Training results

Training Loss Epoch Step Validation Loss Map Map 50 Map 75 Map Small Map Medium Map Large Mar 1 Mar 10 Mar 100 Mar Small Mar Medium Mar Large Map Basketball Mar 100 Basketball Map Player Mar 100 Player Map Referee Mar 100 Referee
No log 1.0 461 1.7649 0.0495 0.1443 0.0205 0.0386 0.0955 0.1299 0.0262 0.1028 0.1539 0.1341 0.2707 0.5936 0.0048 0.0142 0.1345 0.3828 0.0093 0.0646
1.9775 2.0 922 1.5434 0.1063 0.2515 0.0673 0.0772 0.1801 0.4147 0.0497 0.2231 0.2853 0.2337 0.3809 0.6431 0.0004 0.0211 0.2031 0.4458 0.1154 0.389
1.7156 3.0 1383 1.4497 0.139 0.3385 0.0796 0.1042 0.3513 0.4331 0.0722 0.2634 0.3277 0.2719 0.519 0.6221 0.0019 0.042 0.21 0.4515 0.2049 0.4895
1.5817 4.0 1844 1.3623 0.191 0.4015 0.1554 0.1364 0.2989 0.4691 0.1062 0.2979 0.3572 0.3144 0.4426 0.6789 0.002 0.0515 0.2474 0.4896 0.3236 0.5305
1.4789 5.0 2305 1.3544 0.1874 0.4189 0.1338 0.1415 0.357 0.4253 0.1031 0.2946 0.36 0.3174 0.5002 0.6838 0.0032 0.0591 0.2486 0.4917 0.3104 0.5292
1.4235 6.0 2766 1.2867 0.211 0.4444 0.1715 0.154 0.3671 0.298 0.118 0.3084 0.3629 0.3113 0.5314 0.6814 0.0032 0.0494 0.2607 0.4919 0.3692 0.5474
1.3454 7.0 3227 1.3163 0.2024 0.4324 0.1654 0.1381 0.362 0.4507 0.114 0.3076 0.3557 0.3114 0.4828 0.651 0.0052 0.0581 0.2397 0.4762 0.3623 0.5329
1.3253 8.0 3688 1.2467 0.2169 0.4386 0.1888 0.1576 0.3525 0.4086 0.1128 0.327 0.3884 0.344 0.5275 0.6926 0.0105 0.0693 0.2758 0.525 0.3643 0.5708
1.276 9.0 4149 1.2343 0.2242 0.4614 0.1877 0.1659 0.3289 0.3643 0.1273 0.3341 0.3883 0.3438 0.4961 0.6814 0.0094 0.0896 0.2701 0.5176 0.3932 0.5579
1.2626 10.0 4610 1.2377 0.2324 0.4659 0.2072 0.1803 0.3146 0.3717 0.1261 0.3469 0.3967 0.3548 0.4797 0.624 0.0101 0.0849 0.2777 0.5198 0.4094 0.5855
1.2274 11.0 5071 1.2398 0.2358 0.4759 0.2075 0.1795 0.3235 0.4632 0.1338 0.3479 0.3952 0.3457 0.4835 0.6647 0.0128 0.0864 0.2831 0.5323 0.4114 0.5669
1.2026 12.0 5532 1.1964 0.2407 0.4828 0.2133 0.1808 0.3974 0.4632 0.1329 0.3571 0.4064 0.3481 0.5671 0.6966 0.0151 0.0897 0.2866 0.5305 0.4203 0.5991
1.2026 13.0 5993 1.2058 0.2367 0.4879 0.201 0.1919 0.3161 0.4254 0.1314 0.3515 0.398 0.3516 0.4787 0.6789 0.019 0.0999 0.287 0.531 0.404 0.5632
1.1779 14.0 6454 1.1949 0.2365 0.4716 0.2179 0.1759 0.3254 0.4736 0.1268 0.359 0.4064 0.3592 0.5701 0.6495 0.0203 0.1028 0.2945 0.542 0.3946 0.5743
1.1578 15.0 6915 1.2130 0.2291 0.468 0.2024 0.1673 0.3503 0.4148 0.1238 0.3551 0.4049 0.3668 0.485 0.6059 0.02 0.1102 0.2821 0.5349 0.3851 0.5697
1.131 16.0 7376 1.2012 0.2384 0.478 0.2139 0.1848 0.2908 0.4595 0.1314 0.3597 0.4048 0.3616 0.5554 0.6382 0.0174 0.1082 0.291 0.5388 0.4068 0.5672
1.1208 17.0 7837 1.1768 0.2517 0.4928 0.2365 0.2057 0.3507 0.4399 0.1352 0.3697 0.4208 0.38 0.5615 0.6201 0.0222 0.1082 0.3005 0.5574 0.4324 0.5969
1.0992 18.0 8298 1.1605 0.2403 0.4776 0.2194 0.1967 0.2987 0.4002 0.1315 0.3655 0.4148 0.3761 0.5148 0.5838 0.0216 0.1004 0.2958 0.5538 0.4034 0.5901
1.0793 19.0 8759 1.1529 0.2496 0.4954 0.2307 0.2071 0.3352 0.4166 0.133 0.3752 0.4255 0.3868 0.5096 0.6431 0.0278 0.1219 0.3072 0.567 0.4139 0.5877
1.0648 20.0 9220 1.1573 0.2539 0.505 0.2326 0.2033 0.3472 0.4315 0.1389 0.3818 0.4308 0.3896 0.5701 0.6275 0.0327 0.1387 0.3021 0.5606 0.427 0.5931
1.05 21.0 9681 1.1417 0.257 0.505 0.2463 0.2135 0.3753 0.4454 0.1392 0.3862 0.4331 0.3949 0.5359 0.6696 0.0339 0.1388 0.3103 0.5656 0.4267 0.5948
1.0362 22.0 10142 1.1439 0.259 0.5124 0.2466 0.2112 0.3458 0.3431 0.1406 0.3832 0.4307 0.3895 0.4829 0.6402 0.0326 0.1252 0.312 0.5706 0.4324 0.5963
1.0248 23.0 10603 1.1317 0.2641 0.5182 0.2514 0.215 0.3594 0.3094 0.1445 0.3838 0.4319 0.3942 0.5376 0.6137 0.0334 0.1296 0.3123 0.5687 0.4467 0.5972
1.0173 24.0 11064 1.1485 0.2581 0.5057 0.247 0.2102 0.3723 0.4356 0.1414 0.3819 0.4295 0.3906 0.5416 0.6681 0.0334 0.1372 0.3158 0.5696 0.4251 0.5817
1.0082 25.0 11525 1.1344 0.2642 0.5158 0.2495 0.2176 0.3517 0.4473 0.1467 0.3843 0.4322 0.3915 0.554 0.6377 0.0354 0.1386 0.3158 0.5685 0.4414 0.5894
1.0082 26.0 11986 1.1267 0.2648 0.5175 0.2514 0.2147 0.3598 0.4399 0.1489 0.3868 0.4341 0.3942 0.5624 0.6294 0.0381 0.1422 0.3181 0.5706 0.4381 0.5894
1.0006 27.0 12447 1.1296 0.2687 0.5208 0.2581 0.2198 0.3694 0.4415 0.1506 0.3887 0.4359 0.3967 0.5455 0.6333 0.0439 0.1464 0.3188 0.5727 0.4434 0.5885
0.9989 28.0 12908 1.1237 0.2675 0.5191 0.2562 0.2202 0.3769 0.3991 0.1484 0.3897 0.4374 0.397 0.5484 0.6333 0.0411 0.1454 0.3204 0.5735 0.441 0.5932
0.9952 29.0 13369 1.1251 0.2687 0.5204 0.2582 0.221 0.3689 0.3963 0.15 0.3904 0.4385 0.3984 0.5485 0.6314 0.0426 0.1474 0.3207 0.5744 0.4427 0.5936
0.9909 30.0 13830 1.1249 0.2684 0.5212 0.2578 0.22 0.377 0.395 0.1497 0.3903 0.4383 0.398 0.5486 0.6314 0.0431 0.147 0.3203 0.5743 0.4419 0.5936

Framework versions

  • Transformers 4.46.3
  • Pytorch 2.5.1
  • Datasets 3.1.0
  • Tokenizers 0.20.3
Downloads last month
40
Safetensors
Model size
41.6M params
Tensor type
F32
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for clee9/detr_finetuned_30

Finetuned
(456)
this model