Edit model card

Mongolian_GPT_FakeNews_Comprehendo

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

  • Loss: 1.3175
  • Accuracy: 0.8393

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: 3e-05
  • train_batch_size: 16
  • eval_batch_size: 3
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 30

Training results

Training Loss Epoch Step Validation Loss Accuracy
0.1788 1.0 11 0.7540 0.8571
0.2589 2.0 22 0.6399 0.8571
0.1117 3.0 33 0.6795 0.8125
0.0829 4.0 44 0.6606 0.8571
0.0037 5.0 55 0.7375 0.8482
0.0017 6.0 66 0.8388 0.8393
0.0009 7.0 77 0.8872 0.8393
0.0007 8.0 88 0.9371 0.8393
0.0005 9.0 99 0.9949 0.8393
0.0004 10.0 110 1.0329 0.8393
0.0003 11.0 121 1.0626 0.8393
0.0003 12.0 132 1.0800 0.8393
0.0002 13.0 143 1.0993 0.8393
0.0002 14.0 154 1.1330 0.8393
0.0002 15.0 165 1.1925 0.8393
0.0001 16.0 176 1.2286 0.8393
0.0001 17.0 187 1.2468 0.8393
0.0001 18.0 198 1.2586 0.8393
0.0001 19.0 209 1.2686 0.8393
0.0001 20.0 220 1.2758 0.8393
0.0001 21.0 231 1.2836 0.8393
0.0001 22.0 242 1.2914 0.8393
0.0001 23.0 253 1.2978 0.8393
0.0001 24.0 264 1.3027 0.8393
0.0001 25.0 275 1.3070 0.8393
0.0001 26.0 286 1.3106 0.8393
0.0001 27.0 297 1.3131 0.8393
0.0001 28.0 308 1.3156 0.8393
0.0001 29.0 319 1.3170 0.8393
0.0001 30.0 330 1.3175 0.8393

Framework versions

  • Transformers 4.38.2
  • Pytorch 2.2.1+cu121
  • Datasets 2.18.0
  • Tokenizers 0.15.2
Downloads last month
2
Safetensors
Model size
184M params
Tensor type
F32
·

Finetuned from