VOOZH about

URL: https://huggingface.co/merve/rfdetr-docvqa-media3-v2

⇱ merve/rfdetr-docvqa-media3-v2 · Hugging Face

rfdetr-docvqa-media3-v2

This model is a fine-tuned version of Roboflow/rf-detr-base on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 16.2431
  • Map: 0.0037
  • Map 50: 0.0137
  • Map 75: 0.0004
  • Map Small: -1.0
  • Map Medium: 0.0018
  • Map Large: 0.0039
  • Mar 1: 0.0203
  • Mar 10: 0.0793
  • Mar 100: 0.1731
  • Mar Small: -1.0
  • Mar Medium: 0.025
  • Mar Large: 0.1816
  • Map Chart: 0.0046
  • Mar 100 Chart: 0.1474
  • Map Image: 0.0062
  • Mar 100 Image: 0.3269
  • Map Signature: 0.0001
  • Mar 100 Signature: 0.045

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: 42
  • optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_steps: 0.05
  • num_epochs: 10

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 Chart Mar 100 Chart Map Image Mar 100 Image Map Signature Mar 100 Signature
40.0459 1.0 33 25.5239 0.0 0.0 0.0 -1.0 0.0 0.0 0.0 0.0 0.0067 -1.0 0.0 0.007 0.0 0.0 0.0 0.0 0.0 0.02
15.4391 2.0 66 25.5408 0.0 0.0003 0.0 -1.0 0.0 0.0 0.0 0.003 0.003 -1.0 0.0 0.0031 0.0 0.0053 0.0001 0.0038 0.0 0.0
13.9985 3.0 99 22.7635 0.0002 0.0004 0.0002 -1.0 0.0 0.0002 0.007 0.0105 0.0155 -1.0 0.0 0.0158 0.0006 0.0316 0.0 0.0 0.0 0.015
13.1759 4.0 132 21.8405 0.0013 0.0063 0.0 -1.0 0.0 0.0013 0.0035 0.0233 0.0392 -1.0 0.0 0.0412 0.0025 0.0421 0.0012 0.0654 0.0001 0.01
12.6280 5.0 165 21.5377 0.0015 0.0042 0.0011 -1.0 0.0059 0.0015 0.0211 0.0228 0.1041 -1.0 0.125 0.1001 0.0037 0.2 0.0008 0.0923 0.0 0.02
12.3205 6.0 198 19.0510 0.0012 0.0046 0.0004 -1.0 0.0042 0.0012 0.0056 0.0642 0.1349 -1.0 0.025 0.1397 0.0019 0.1579 0.0015 0.1769 0.0001 0.07
10.5714 7.0 231 16.7398 0.0023 0.01 0.0006 -1.0 0.0134 0.0023 0.0158 0.0647 0.1626 -1.0 0.15 0.1635 0.0049 0.1474 0.0021 0.3154 0.0001 0.025
10.2309 8.0 264 16.9129 0.0135 0.0575 0.0006 -1.0 0.0072 0.0145 0.0126 0.0607 0.1619 -1.0 0.075 0.1676 0.0125 0.1053 0.028 0.3154 0.0001 0.065
9.5882 9.0 297 16.5097 0.0138 0.0327 0.0006 -1.0 0.0007 0.0142 0.0257 0.0652 0.1964 -1.0 0.05 0.2046 0.034 0.1737 0.0073 0.3654 0.0001 0.05
9.5828 10.0 330 16.2431 0.0037 0.0137 0.0004 -1.0 0.0018 0.0039 0.0203 0.0793 0.1731 -1.0 0.025 0.1816 0.0046 0.1474 0.0062 0.3269 0.0001 0.045

Framework versions

  • Transformers 5.12.1
  • Pytorch 2.12.0+cu130
  • Datasets 5.0.0
  • Tokenizers 0.22.2
Downloads last month
-
Safetensors
Model size
31.9M params
Tensor type
F32
·

Model tree for merve/rfdetr-docvqa-media3-v2

Finetuned
(2)
this model