Hugging Face's logo

niujiahui
/
detr-finetuned-cppe-5-10k-steps

Object Detection
Transformers
TensorBoard
Safetensors
detr
vision
Generated from Trainer
Model card Files
xet
 Metrics Community
detr-finetuned-cppe-5-10k-steps / README.md
niujiahui's picture
niujiahui
End of training
f494f6e verified
preview code
|
raw
history blame contribute delete
5.9 kB
metadata 
license: apache-2.0
base_model: facebook/detr-resnet-50
tags:
  - object-detection
  - vision
  - generated_from_trainer
model-index:
  - name: detr-finetuned-cppe-5-10k-steps
    results: []

detr-finetuned-cppe-5-10k-steps

This model is a fine-tuned version of facebook/detr-resnet-50 on the cppe-5 dataset. It achieves the following results on the evaluation set:

  • Loss: 1.6932
  • Map: 0.1289
  • Map 50: 0.2606
  • Map 75: 0.1098
  • Map Small: 0.0355
  • Map Medium: 0.1082
  • Map Large: 0.1676
  • Mar 1: 0.1429
  • Mar 10: 0.2628
  • Mar 100: 0.2869
  • Mar Small: 0.1256
  • Mar Medium: 0.2299
  • Mar Large: 0.3635
  • Map Coverall: 0.399
  • Mar 100 Coverall: 0.6383
  • Map Face Shield: 0.0257
  • Mar 100 Face Shield: 0.1557
  • Map Gloves: 0.0535
  • Mar 100 Gloves: 0.2772
  • Map Goggles: 0.0002
  • Mar 100 Goggles: 0.0031
  • Map Mask: 0.166
  • Mar 100 Mask: 0.36

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: 8
  • eval_batch_size: 8
  • seed: 1337
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 10.0
  • mixed_precision_training: Native AMP

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 Coverall Mar 100 Coverall Map Face Shield Mar 100 Face Shield Map Gloves Mar 100 Gloves Map Goggles Mar 100 Goggles Map Mask Mar 100 Mask
2.6856 1.0 107 2.4604 0.0152 0.0396 0.0096 0.0051 0.0048 0.0234 0.0535 0.1109 0.1286 0.0401 0.089 0.1752 0.0622 0.3509 0.0 0.0 0.0045 0.1174 0.0 0.0 0.0092 0.1747
2.1242 2.0 214 2.1711 0.0464 0.1115 0.033 0.01 0.0466 0.0621 0.0776 0.1595 0.1801 0.0647 0.1406 0.2105 0.1861 0.5176 0.0 0.0 0.0139 0.1437 0.0 0.0 0.0317 0.2391
1.9759 3.0 321 2.0671 0.0662 0.1477 0.053 0.0129 0.0709 0.0846 0.0964 0.1814 0.199 0.057 0.1639 0.2279 0.2207 0.5608 0.0 0.0 0.0222 0.1688 0.0 0.0 0.0881 0.2653
1.8435 4.0 428 1.9923 0.084 0.1759 0.0717 0.0156 0.0673 0.1068 0.1017 0.1929 0.2119 0.0717 0.1646 0.2621 0.2915 0.5856 0.0 0.0 0.0302 0.1835 0.0 0.0 0.0982 0.2907
1.7693 5.0 535 1.9163 0.0851 0.181 0.0718 0.0201 0.0721 0.1072 0.1005 0.1956 0.2137 0.0924 0.1684 0.2556 0.2895 0.5559 0.004 0.0063 0.0256 0.1808 0.0 0.0 0.1064 0.3253
1.6961 6.0 642 1.8520 0.1045 0.2193 0.0909 0.0529 0.0938 0.133 0.1183 0.2171 0.2416 0.1111 0.2093 0.2975 0.3181 0.5748 0.0062 0.0329 0.0432 0.2598 0.0 0.0 0.1549 0.3404
1.6116 7.0 749 1.7836 0.1118 0.2368 0.089 0.0334 0.0935 0.1543 0.1308 0.2439 0.2684 0.1294 0.2151 0.3409 0.3489 0.6059 0.0081 0.1165 0.0517 0.2888 0.0 0.0 0.1503 0.3307
1.5518 8.0 856 1.7223 0.1235 0.2558 0.1096 0.0336 0.1039 0.1553 0.135 0.2555 0.2765 0.1211 0.2311 0.3411 0.3847 0.6203 0.0237 0.1494 0.0556 0.2661 0.0001 0.0015 0.1537 0.3453
1.5112 9.0 963 1.6986 0.1268 0.2639 0.1029 0.0318 0.1025 0.1676 0.1444 0.2626 0.2864 0.1204 0.2328 0.3611 0.3928 0.6392 0.0274 0.157 0.0536 0.2741 0.0012 0.0062 0.1591 0.3556
1.4924 10.0 1070 1.6932 0.1289 0.2606 0.1098 0.0355 0.1082 0.1676 0.1429 0.2628 0.2869 0.1256 0.2299 0.3635 0.399 0.6383 0.0257 0.1557 0.0535 0.2772 0.0002 0.0031 0.166 0.36

Framework versions

  • Transformers 4.42.4
  • Pytorch 1.13.1+cu117
  • Datasets 2.20.0
  • Tokenizers 0.19.1